You are on page 1of 751



This page intentionally left blank

Department of Microbiology
University of Massachusetts
Amherst, Massachusetts


Washington, DC
Address editorial correspondence to ASM Press, 1752 N St. NW, Washington, DC
20036-2904, USA

Send orders to: ASM Press, P.O. Box 605, Herndon, VA 20172, USA
Phone: (800) 546-2416 or (703) 661-1593
Fax: (703) 661-1501

Copyright 2010 ASM Press

American Society for Microbiology
1752 N St. NW
Washington, DC 20036-2904

Library of Congress Cataloging-in-Publication Data

Norkin, Leonard C.
Virology : molecular biology and pathogenesis / Leonard C. Norkin.
p. ; cm.
Includes bibliographical references and index.
ISBN 978-1-55581-453-3 (hardcover : alk. paper) 1. VirologyTextbooks.
2. Molecular virologyTextbooks. 3. Virus diseases
PathogenesisTextbooks. I. Title.
[DNLM: 1. Virusespathogenicity. 2. Genome, Viral. 3. Virus Diseasesetiology.
4. Virusesgenetics. QW 160 N841v 2010]
QR360.N67 2010

10 9 8 7 6 5 4 3 2 1

All rights reserved

Printed in Canada

Cover and interior design: Susan Brown Schmidler

Illustrations: Lineworks, Inc.

Cover illustration: Structure and molecular organization of a Sindbis virus particle.

Sindbis virus is a member of the togavirus family of enveloped plus-strand RNA viruses.
The surface features of the particle were determined by cryo-electron microscopy,
which yielded hundreds of highly detailed, two-dimensional images, from which
a three-dimensional image was generated using powerful computer programs.
A cross-section through the particle, showing the envelope glycoproteins
(blue), the lipid bilayer (green), the nucleocapsid (red), the mixed RNA-protein
region (orange), and the genomic plus-strand RNA (magenta), is superimposed on the
three-dimensional image. Protein structures were solved by X-ray crystallography,
and then fitted into the cryo-EM structure. See Figure 8.1 in the book for the complete
image. Adapted from W. Zhang et al., J. Virol. 76:1164511648, 2002, with permission.
I dedicate this book to my wife, Arline; my sons, Dave and Mike, and their wives,
Mina and Debbie; and my grandchildren, Luke, Maya, and Theo.

Human subtlety will never devise an invention more beautiful, more simple
or more direct than does Nature, because in her inventions, nothing is lacking
and nothing is superfluous.
Leonardo da Vinci (14521519)
This page intentionally left blank

Preface xix

PART I Introduction 1
1 A Selective History on the Nature of Viruses 3
Introduction 3
The Early Years: Discoverers and Pioneers 4
The First Stirrings of the Molecular Era 7
The Phage Group 10
Phage Growth: Eclipse and Replication 11
Defining Viruses 15
Are Viruses Alive? 16
Origin of Viruses 17
The Modern Era of Animal Virology 17

2 Biosynthesis of Viruses: an Introduction to Virus

Classification 20
T-Even Bacteriophages as a Model System 20
T-Even Phage Structure and Entry 22
Sequence of Phage Biosynthetic Events 24
Phage Protein Synthesis 24
RNA Metabolism in Infected Cells 25
Assembly of Progeny Phages 27
Packaging DNA within the Phage Particle 28
Unique Features of T-Even Phages 30
Modified Bases 30
Regulated Gene Expression 30
Phage Release: Lysozyme and the rII Region 32
Bacteriophage Lambda (): Lysogeny and Transduction 33
Some Final Comments on Bacteriophages 39

Introduction to the Animal Viruses 39

Animal Virus Structure 39
Entry of Animal Viruses 42
The Families of Animal Viruses: Principles of Classification 45
Viral Genetic Systems: the Baltimore Classification Scheme 45

3 Modes of Virus Infection and Disease 50

Introduction 50
Portals of Entry 50
Routes of Dissemination 54
Hematogenous and Neural Dissemination 54
The Placenta and the Fetus 60
Acute versus Persistent Infections 63
Acute Infections 64
Persistent Infections 66
Slow Infections 66
Chronic Infections 68
Latent Infections 74
Transmissible Spongiform Encephalopathies: Prions 76

4 Host Defenses and Viral Countermeasures 80

Introduction 80
Overview of Defenses 81
Physical Barriers against Infection 82
The Innate Immune System 82
Cytokines: the IFNs 84
Cytokines: TNF-, Some Other Cytokines, and Inflammation 88
Macrophages, Neutrophils, NK Cells, and Antibody-Dependent
Cellular Cytotoxicity 89
The Complement System 95
Viral Evasion of Innate Immunity 98
Evasion of IFNs 98
Evasion of Cytokines 100
Evasion of NK Cells and ADCC 101
Evasion of Complement 102
APOBEC3G and the HIV Vif Protein 103
The Adaptive Immune System 103
Antibodies and B Cells 104
Antibody Diversity 106
Viral Evasion of Antibodies 110
Cell-Mediated Immunity 113
Antigen Presentation by MHC Class I Molecules 119
Antigen Presentation by MHC Class II Molecules 119
The Rationale for MHC Restriction 121
Activation of Th Cells: Dendritic Cells and B Cells 124
Activation of B Cells 126
Activation of CTLs 128
Mechanism of Action of CTLs 129

T Cells and Antiviral Cytokines 131

Viral Evasion of Cell-Mediated Immunity 132
Inhibition of Antigen Presentation to CTLs 132
Inhibition of Antigen Presentation to Helper T Cells 134
Inhibition of Apoptosis 134
Immunological Memory 136
Self Tolerance 137
The Immune System in Disease 140
Immunopathology 140
Autoimmune Disease 141

PART II Virus Replication and Pathogenesis 147

RNA Viruses: Double Stranded 149

5 Reoviruses 149
Introduction 149
Structure, Binding, Entry, and Uncoating 150
Reovirus Binding and Entry into the Cell 150
Structure, Uncoating, and Entry into the Cytoplasm 154
The Reovirus Genome: Transcription and Translation 156
The Particle-Associated RNA Polymerase 156
The Segmented Reovirus Genome 158
Conversion of ISVPs to Cores 161
Replication and Encapsidation of the Reovirus Genome 165
Synthesis of Double-Stranded RNA 165
Assembly of Progeny Subviral Particles and Encapsidation of RNA Segments 166
Reoviruses and IFN 166
Primary versus Secondary Transcription 167
Final Virus Assembly 168
Pathogenesis 168
Orthoreoviruses 168
Rotaviruses 168
Coltiviruses 169

RNA Viruses: Single Stranded 171

6 Picornaviruses 171
Introduction 171
Structure, Binding, and Entry 172
Picornavirus Structure 172
Rhinovirus Receptor and Binding: the Canyon Hypothesis 173
The Poliovirus Receptor 177
Receptors for Coxsackieviruses and Other Enteroviruses 179
Receptors for FMDVs 180
Poliovirus and Rhinovirus Entry: Some General Points 181
Poliovirus Entry 181

Human Rhinovirus Entry 183

Poliovirus and Rhinovirus Entry: Why the Differences? 184
Translation 184
Translation: Part I 184
Translation: a Digression 187
The RNA Phages 187
Picornaviruses versus RNA Phages: Why the Differences? 191
Translation: Part II 192
Transcription and Genome Replication 194
Assembly and Maturation 198
Medical Aspects 200
Poliovirus 200
Rhinoviruses: the Common Cold 203
Coxsackievirus and Echovirus 205
Viral Hepatitis: Hepatitis A 205

7 Flaviviruses 207
Introduction 207
Structure and Entry 208
Replication 209
Assembly and Release 209
Historic Interlude: Identification of Hepatitis C Virus 212
West Nile Virus: an Emerging Virus 213
Epidemiology and Pathogenesis 214
General Principles of Arthropod Transmission 214
Infection, Dissemination, and Determinants of Pathogenesis 215
Hemorrhagic Fever Viruses: Yellow Fever and Dengue Viruses 217
Encephalitis Viruses: Japanese Encephalitis, St. Louis Encephalitis,
and West Nile Viruses 218
Japanese Encephalitis and St. Louis Encephalitis Viruses 218
West Nile Virus 220
Hepatitis C Virus 220

8 Togaviruses 224
Introduction 224
Structure and Entry 225
Transcription, Translation, and Genome Replication 228
Assembly and Maturation 231
Epidemiology and Pathogenesis 232
Alphaviruses That Cause Encephalitis: Eastern, Western, and
Venezuelan Equine Encephalitis Viruses 232
Alphaviruses That Cause Arthritis: Chikungunya, Ross River, and Sindbis Viruses 232
Rubella Virus 232

9 Coronaviruses 235
Introduction 235
Structure 236
Entry 237
Genome Organization and Expression 238
Coronavirus mRNAs and Their Translation 238
Coronavirus Transcription 243
Coronavirus Recombination 247
Coronavirus Reverse Genetics 248
Assembly and Release 250
Medical Aspects 251
SARS 252

10 Rhabdoviruses 261
Introduction 261
Structure 262
Entry 263
Genome Organization, Expression, and Replication 264
The General Transcriptional Strategy of Viruses That Contain Negative-Sense
RNA Genomes 264
Gene Organization and Transcription 264
Replication 267
Assembly and Release 268
Medical Aspects 269
Cytopathic Effects 269
VSV 270
Rabies Virus 270

11 Paramyxoviruses 273
Introduction 273
Structure 274
Entry 276
Syncytium Formation 279
Genome Organization, Expression, and Replication 279
Genome Organization and Transcription 279
Replication 281
Assembly and Release 282
Medical Aspects 282
Measles 282
Clinical Conditions 282
SSPE 285
Where Did Measles Come from? 288
Measles Vaccine 289
Mumps 290
Parainfluenza Viruses 290
Respiratory Syncytial Virus 291

HMPV 292
Hendra, Nipah, and Menangle Viruses 294
CDV 295

12 Orthomyxoviruses 296
Introduction 296
Structure 298
Entry 299
Endocytosis and Intracellular Trafficking of Influenza Virus-Containing Endosomes 300
Viral Membrane Fusion: the HA Protein 302
Release of the Genome from the Envelope: the M2 Protein 307
Transport of vRNPs to the Cell Nucleus 308
Genome Organization, Transcription, and Replication 311
Genome Organization and Transcription 311
Why Does Influenza Virus Transcription Really Occur in the Nucleus? 317
Replication 317
The NS1 and PB1-F2 Proteins 318
Assembly and Release 322
Assembly of vRNPs 322
Two Issues Regarding Intracellular Targeting of Viral Components 322
Transport of vRNPs from Nucleus to Cytoplasm 322
Apical Targeting of Virus Components and Final Virus Assembly 323
Encapsidating Eight Distinct Genomic Segments 325
Release: the Influenza Virus NA Protein 325
Medical Aspects 325
Pathology and Clinical Syndromes 325
The Flu Pandemic of 1918 326
Origin of Epidemic and Pandemic Strains: Antigenic Drift and Antigenic Shift 328
Virulence of the 1918 Pandemic Influenza Virus: Opening Pandoras Box 333
Influenza Drugs and Vaccines 336
Avian Influenza and Humans 337
Preparedness for an Outbreak of Avian Influenza 341

13 Miscellaneous RNA Viruses 346

Introduction 346
Arenaviruses 347
Lassa Fever Virus: an Early Emerging Virus 348
LCMV and Immunopathology 348
The Ambisense Arenavirus Genome 349
Bunyaviruses 350
Sin Nombre and Hantaan Viruses 351
The Tripartite Bunyavirus Genome 351
Bornaviruses 352
BDV and Human Neuropsychiatric Disease 352
Bornavirus Genomes and Their Nuclear Transcription and Replication 354
Caliciviruses 355
Norwalk Virus (Norovirus) 355
Calicivirus Replication 356
Astroviruses 356

Hepatitis E-Like Viruses 357

Filoviruses 358
Structure and Replication 358
Marburg and Ebola Viruses 358
Is Ebola Virus the Andromeda Strain? 361

DNA Viruses: Single Stranded 362

14 Parvoviruses 362
Introduction 362
Parvovirus Structure, Binding, and Entry 363
Parvovirus Genomes 365
Replication of Parvovirus Single-Stranded DNA 367
A General Model: the Rolling Hairpin 368
Encapsidation of Progeny DNA 371
Genome Organization and Expression 374
The Dependovirus Life Cycle 374
Integration of Dependovirus Genomes 375
Medical Aspects 376

DNA Viruses: Double Stranded 378

15 Polyomaviruses 378
Introduction 378
Structure 380
Entry and Uncoating 382
Genome Organization and Expression 385
Genome Organization 385
The Regulatory Region 386
T Antigen in Temporal Regulation of Transcription 390
Splicing Pattern of Viral mRNA 394
DNA Replication 395
Late Proteins and Assembly 399
T Antigens and Neoplasia 401
SV40 in Its Natural Host 410
SV40 in Humans 411
JCV and BKV in Humans 414
KI, WU, and Merkel Cell Human Polyomaviruses 415
SV40-Based Gene Delivery Vectors 418

16 Papillomaviruses 419
Introduction 419
Structure 420
Entry and Uncoating 421

Genome Organization and Expression 424

Replication 429
Release and Transmission 432
Cell Transformation and Oncogenesis: Cervical Carcinoma 433
Molecular Mechanisms 433
Stages of Cancer Development 439
Tissue Microenvironment 439
Treatment and Prevention 441
HPV Vaccine To Prevent Cervical Cancer 441
Oropharyngeal Cancer 442

17 Adenoviruses 444
Introduction 444
Structure 446
Entry and Uncoating 448
Genome Organization and Expression 452
E1A and E1B Proteins in Replication and Neoplasia 456
Shutoff of Host Protein Synthesis 458
DNA Replication 460
Assembly 462
Evasion of Host Defenses 464
Clinical Syndromes 465
Recombinant Adenoviruses 466
Gene Therapy 466
Cancer Therapies 467
Vaccines: HIV and Influenza Virus 469

18 Herpesviruses 471
Introduction 471
Structure 473
Entry and Uncoating 475
Genome Organization and Expression 481
DNA Replication 487
Assembly, Maturation, and Release 488
Assembly of Procapsids 488
Encapsidation of DNA 490
Final Assembly and Release 491
Latent Infection and Immune Evasion 493
HSV 494
EBV 499
HCMV 503
Clinical Syndromes 506
HSV 507
EBV 508
HCMV 512

VZV 513
HHV-6 514
HHV-7 515
HHV-8 515
Treatment and Prevention 516
Antiviral Drugs 516
Vaccines 519

19 Poxviruses 521
Introduction 521
Structure 523
Entry and Uncoating 525
IMVs 526
EVs 528
Genome Organization and Expression 529
DNA Replication 533
Assembly, Maturation, and Release 534
Immune Evasion 538
Countermeasures against Innate Defenses 538
Countermeasures against Adaptive Defenses 541
Clinical Syndromes 543
Smallpox 543
Molluscum Contagiosum 544
Vaccinia Virus and the Smallpox Vaccine 546
Smallpox and Bioterrorism 547

PART III Retroviruses: Oncogenes and Cancer,

HIV and AIDS 551
20 The Retroviruses: the RNA Tumor Viruses 553
Introduction 553
A Historical Interlude 556
Structure 559
Entry 561
The Replication Cycle: a Brief Overview 562
Genome Organization 563
Reverse Transcription 564
Integration 568
Gene Expression and Replication 569
Assembly, Maturation, and Release 572
Gag and Gag-Pol 572
Env and Budding 575
Oncogenesis 576
Oncogene-Transducing Oncogenic Retroviruses 577
Oncogene-Deficient Oncogenic Retroviruses 588

Retrovirus Vectors for Gene Therapy 591

Spumaviruses 592
Retroelements 593

21 The Retroviruses: Lentiviruses, Human Immunodeficiency

Virus, and AIDS 597
Introduction 597
Historical Background and Social Issues 597
HIV and the Lentiviruses 600
Discovery of HIV as the Cause of AIDS 602
Structure 605
Routes of Transmission and Mechanisms of Infection 606
Routes of Transmission 606
Entry into the Host 607
Entry into Cells 610
The HIV Coreceptors 611
Entry into Human Placental Trophoblasts: a Special Case? 614
Therapeutic Strategies That Target Entry 615
Genome Organization, Expression, and Replication 616
Genome Organization and Expression 616
The Regulatory Proteins: Tat and Rev 619
Assembly, Maturation, and Budding 621
HIV Assembly in Macrophages: Macrophages in HIV Pathogenesis 622
The Course of HIV/AIDS 623
The Acute HIV Disease Stage 623
The Asymptomatic HIV Disease Stage 624
The Symptomatic HIV Disease Stage 627
AIDS: Advanced HIV Disease Stage 627
The Viral Set Point and HAART 627
Immunopathogenesis 629
Immune Evasion and Manipulation 634
The Accessory Proteins: Nef, Vpu, Vif, Vpr, and Vpx 635
Tat Revisited 639
Evasion and Subversion of Antibody-Mediated Immunity 640
HIV and Complement 640
HIV and the IFNs 641
HIV (SIV) in Its Natural Host and Its Origins as a Human Virus 641
Clinical Manifestations of HIV/AIDS 643
Therapies and Vaccines 647
Antiretroviral Drugs 647
Anti-HIV Vaccines 653
Prevention 655
Testing for HIV 657
HIV/AIDS in Sub-Saharan Africa 662
The General Problem 662
The Case of South Africa 665
The Human T-Cell Leukemia (Lymphotropic) Viruses 666

22 Hepadnaviruses and Hepatitis Delta Virus 672

The Hepadnaviruses: DNA Retroviruses 672
Introduction 672
Structure 674
Replication 675
Entry 675
Genome Organization 676
Gene Expression 678
Reverse Transcription 679
The Foamy Viruses: Reprise 684
Maturation and Release 684
Routes of Transmission and Pathogenesis 685
Hepatocellular Carcinoma 688
HBV Vaccine 689
HBV Therapies 689
Hepatitis Delta Virus: a Subviral Pathogen 690
Introduction 690
Replication 690
Routes of Transmission and Pathogenesis 693

Index 695
This page intentionally left blank

Virology: Molecular Biology and Pathogenesis is meant to be used as a textbook for a compre-
hensive virology course aimed at advanced undergraduate and graduate students. With a
handful of worthy virology textbooks already available, what justification might there be for
yet another one, moreover one written by a single author? There are three reasons. First, and
most importantly, Virology: Molecular Biology and Pathogenesis was conceived and organized
to express my avid belief that the best way to teach virology is to discuss viruses in the context
of virus families. Second, the individual virus families are viewed here within the context of
the Baltimore classification system, a key unifying theme of the book and an approach that
enables students to assume basic facts about the replication strategy of any virus on the basis
of the nature of its genome (e.g., double-stranded DNA, plus-sense versus minus-sense single-
stranded RNA, etc.). I know from more than three decades of teaching virology in this way that
this is a sure-fire strategy for preparing students to approach the journal literature on any virus
intelligently, key relevant knowledge already having been mastered. For the same reason, this
book should continue to serve as a valuable reference for those who have studied from it.
Third, I believe that this volume is unique in the uniform organization of its individual chap-
ters and in its consistent writing style. Thus, Virology: Molecular Biology and Pathogenesis is
intended to be a principles approach to virology, in which the same fundamental principles
serve as the focal points of each chapter. In addition, the individual chapters constitute a con-
tinuous narrative in which key principles recur and are reinforced in different contexts.
Finally, the writing style is often deliberately conversational so as to be more accessible and, I
hope, engaging while not sacrificing rigor.
I began writing Virology: Molecular Biology and Pathogenesis in September 2003, and it has
been the major focus of my professional life ever since. Yet its seeds may have been planted
more than 35 years ago. My doctoral training in the mid-1960s was in the area of bacterial
genetics. After earning my Ph.D. in 1969, I spent two years doing postdoctoral research on the
mouse polyomavirus, a small DNA tumor virus. Suddenly I found myself an assistant profes-
sor at the University of Massachusetts, where I was expected to teach a virology course to se-
nior microbiology majors and graduate students. I was an expert bacterial geneticist and could
have readily taught a graduate course in microbial genetics. Although I had become familiar
with the tumor viruses during my postdoctoral stint, I was far from being an expert animal
virologist. Adding to my dilemma, the virology textbooks of the day were for the most part
descriptive. I was thus at a loss as to how to deliver three 50-minute virology lectures per week
to advanced undergraduate and graduate students. Fortuitously, I came across a review article
by David Baltimore in which he put forward what is referred to as the Baltimore classification


system (Baltimore, 1971). In brief, the Baltimore classification system is based on the different
strategies used by viruses to express their genomes, and importantly, it recognizes that a given
viral genome, by its nature, determines its expression strategy. Moreover, there were only six
classes, or basic strategies of viral genome expression, in the original Baltimore scheme, and
each of the numerous classic animal virus families fits into one of those classes. I had my an-
swer regarding how to teach virology. I essentially followed Baltimores review article and lec-
tured from the original journal articles that were referenced therein. I continued in this way for
the next several years, updating my lectures with more recent journal articles. Knowledge in
the field was increasing much too rapidly, though, for me to go on in this way. I needed to turn
to a textbook. I used several different ones over the years, but for one reason or another, none
was the ideal teaching aid I was seeking. During this time I fantasized about writing a book
that might present the field of virology as I envisioned it, and the current text is a fulfillment
of that vision.
When I first taught virology in 1972, armed with the Baltimore classification system as my
guide, animal virology was in its infancy. Yet it still was a challenge to cover all the pertinent
material that was then available in the forty or so lectures that were allotted. Developments in
the molecular biology of the animal viruses since the early 1970s have been explosive. Further-
more, viral structural biology and virus-cell and virus-host interactions, which were barely
discussed in earlier decades, are now major areas of modern virology. In 1972 there was no
HIV/AIDS, Ebola virus, West Nile virus, SARS, or bird flu to discuss, nor was there a vaccine
to protect women from papovavirus-induced cervical cancer, and what was known of viral
pathogenesis was, for the most part, descriptive. Yet, despite all the developments of the last
few decades, I still have the same forty or so lectures in which to cover virology. How does
one include it all in a single semester course? How does a textbook writer encompass it in
one book? The answer is that one cannot include all of virology between the covers of even
a large tome. One can ask, however, what readers need to know right now that will enable
them to approach the journal literature successfully on any particular virus. Bearing that
standard in mind, the crucial portion of each chapter describes in detail the organization of
the viral genome and its pattern of expression. Bracketing that key topic are viral structure
and entry, followed by assembly, release, and medical issues. Considering the vast amount
of material covered in Virology: Molecular Biology and Pathogenesis, I suspect that most
instructors will not ask their students to master it all. The portions of each chapter that
concern the organization of the viral genome and its strategy of expression, however, are
Virology: Molecular Biology and Pathogenesis begins with two introductory chapters that
recount the discovery of viruses and the recognition of their unique nature. Since the T-even
bacteriophages and bacteriophage played major roles in the development of virology, these
prokaryotic viruses are considered in some detail. While this book is primarily about the ani-
mal viruses, these and other bacteriophages are also discussed in later chapters, at points where
it is informative to compare their unique properties to those of particular animal viruses. For
example, in Chapter 6 the pattern of gene expression of the RNA phages is compared to that
of the picornaviruses for the purpose of better appreciating the reasons underlying the unique
aspects of each.
In the second of the two introductory chapters, we are introduced to the principles of
animal virus classification, which define the classic animal virus families. We then consider
the Baltimore classification scheme, which serves as a unifying principle for much of the
remainder of the book. Before coming to the chapters dedicated to each of the major virus
families, the student encounters chapter 3, which covers the various modes of virus infec-
tion and disease. This chapter provides background and perspective on the specific examples
discussed in the chapters dedicated to the individual virus families. Chapter 4 then considers
host defenses against microorganisms, on one hand, and viral countermeasures to subvert
those host defenses on the other. I hope that students will read this chapter in its entirety,

without necessarily studying it. Instead, I wish for readers to come away with a sense of just
how potent our immune defenses are and yet how astonishingly efficiently evolution has
acted to give rise to viral countermeasures against those host defenses. Without this con-
tinuing evolutionary dance between pathogen and host, there would be neither host nor
virus. Readers should refer back to this chapter when viral immune evasion strategies are
recounted in the contexts of the individual virus families, which we come to next, beginning
with Chapter 5.
An important theme throughout the book is the manner in which some of the most dra-
matic breakthroughs in molecular and cellular biology sprang directly from the study of
viruses. In fact, the field of molecular biology itself owes much to the earlier discovery of
viruses. These early developments are woven into the history of virology, as recounted in
Chapters 1 and 2; subsequent key virus-related breakthroughs are acknowledged in later chap-
ters. Moreover, studies of viruses have led to crucial insights into the molecular basis of cancer,
detailed here mainly in Chapters 15 and 20 but in other chapters as well.
Although Virology: Molecular Biology and Pathogenesis considers medical aspects of virol-
ogy in somewhat more depth than most general texts, it is not meant to be a medical virology
text per se. It does not consider diagnosis, nor is it meant to serve as a guide to therapies for the
practicing physician. Rather, it is meant to give the student, the medical practitioner, and the
basic scientist an understanding of fundamental virology as it relates to viral infection and
Throughout the text there are boxes, which serve several purposes. Some of them review
fundamental molecular and cellular biology that is relevant to the virology being discussed.
Virology is also a human story, however, and I believe that as we learn the details of virology,
it is important that we remain aware of the fact that we owe our knowledge to the efforts of
individual scientists and physicians who at times showed exceptional brilliance and insight.
Some of these individuals ran up against entrenched prevailing dogma and endured continu-
ing humiliation. But they nevertheless persevered, until they were vindicated in the end. Still
others displayed extraordinary physical courage, putting their own lives at risk for the greater
public good. The stories of some of these researchers and clinicians are recounted in the boxes,
and in some cases, in the main text. In addition, I have tried to convey the scientific climate of
the times when some of the earlier developments took place. Perhaps the most important
point of these asides is to help the student develop an appreciation of how the science of virol-
ogy got to where it is todaya fascinating story in its own right. Also, since viruses are agents
of life-threatening infectious diseases, virology, more than most biological sciences, impacts
society in important ways. Additional boxes, therefore, address the serious public policy issues
raised by viral outbreaks and epidemics, vaccines, limited resources, and so forth. Finally,
throughout the text I have tried to highlight key questions that remain to be answered as well
as challenge readers to suggest experimental approaches that might provide these answers.
Other thought-provoking questions are scattered throughout the text simply to better engage
students in the topic at hand.
The choice of specifics to include in Virology: Molecular Biology and Pathogenesis mainly
reflected my own particular interests. But the ASM Press secured for me expert reviewers,
who often steered me towards what is actually of greatest interest to those in each particular
field. For one impulsive (or foolhardy) enough to single-handedly write a comprehensive
textbook on virology, that expert advice has been invaluable; the efforts of these reviewers
are more properly acknowledged below. I add the usual proviso that any errors of interpre-
tation or emphasis are mine alone. I hope that users of this book will bring to my attention
any errors that need correction, as well as any other advice that might make for an improved
second edition.
I have not laced this book with many journal references, and those relatively few (for a com-
prehensive text) that are included may seem idiosyncratic to expert readers. Some articles are refer-
enced for their technical innovations, while others were chosen for their historical interest.
xxii PREF A CE

Yet others were chosen for having been current at the time a particular passage was being writ-
ten and for having suggested an exciting new paradigm. In these instances, only time will tell
whether the new ideas they put forward will become widely accepted.
I justify this rather sparse approach to citations largely because we live today in the age of
the Internet, where up-to-date information is instantly available, and because the rate of sci-
entific progress is so rapid that any fixed set of references will be out of date virtually immedi-
ately. Thus, I urge that you, the reader, study this book with your computer at hand. When you
are seeking clarification or want to know the latest information on any topic, my recommen-
dations are as follows. While there are many Web sites that might turn up the information
sought, the best, most reliable resource is the PubMed site at the U.S. National Library of
Medicine. Go to the site, bookmark it, and use it often. You can search by topic or author, and
the pertinent original journal articles will be listed for you, from most to least current. You will
be able to view the abstract of any paper that appears on the list turned up by your search. In
some cases, you will be able to access complete journal articles either directly on the site or
through arrangements provided by your institution.
You will add immeasurably to your learning experience if you consult the journal literature
as you read this book. Those readers who are not yet experienced at consulting the journal
literature should be forewarned that it is a skill that comes with practice. Do not be discouraged
if the going is rough at first. Here are some practical hints that may help. Begin by scanning the
articles Introduction to determine whether it in fact contains, or might direct you to, the in-
formation you are seeking. Scan the Discussion as well, since it should tell you how the current
paper fits into the bigger picture while it also directs you to additional references that might
contain what you have been seeking. (Scanning is a valuable skill worth acquiring. Impor-
tantly, when scanning, do not stop for details that you may not understand. Keep scanning.)
After scanning the Introduction and Discussion, if you have located a paper that has the an-
swer to your specific question, or if the paper at hand is interesting to you, go back and try to
read it for understanding. Most readers may not yet need to understand experimental details,
but I urge you to try to understand the underlying logic of the experiments and how the data
lead to the authors conclusions. My greatest wish is that this book will serve as a foundation
that will ease you into reading the virology literature, from which you will build a continuing
understanding of viruses and their diseases.
As noted above, each of the chapters of Virology: Molecular Biology and Pathogenesis was
read by one or more reviewers, each of whom provided expert commentary on an individual
chapter. Other scientists, who have had much experience teaching the subject of virology, re-
viewed multiple chapters and, in some instances, the entire book in order to offer advice regard-
ing its merits as a teaching tool. I am grateful to these colleagues for their time and expertise and
for their special insights. As I noted above, a text of this scope by one author would not have
been possible were it not for their advice. In addition, I am most grateful to those reviewers who
also offered kind words of encouragement, which invariably sent me happily back to the mill-
stone with renewed vigor. So, to each of the reviewers enumerated below I add my most sincere
In addition, I am grateful to Eva Szomolanyi-Tsuda (University of Massachusetts Medical
School). A presentation that she gave at a scientific conference helped me to rethink how to
present some of the material in Chapter 4. I am also grateful to Greg Payne, Senior Acquisi-
tions Editor at ASM Press. Greg took an early interest in my book proposal and guided me
through the process of securing a book contract with ASM Press. I have greatly enjoyed my
interactions with him. More recently, I have enjoyed working with John Bell, Senior Produc-
tion Editor at ASM Press. John gently kept me on track towards bringing this project to a
conclusion while remaining sensitive to other demands on my time. Also, I am grateful to
Michael Norkin, who helped me a few times when I was not confident that my own writing
skills were up to the task, and to my student and friend, Dmitry Kuksin, who helped compile
the figures and who provided general computer expertise.
PREF A CE xxiii

Lee Abrahamsen, Bates College Stanley Perlman, University of Iowa
Larry Anderson, Centers for Disease Control and Douglas Richman, University of California at
Prevention San Diego
Walter Atwood, Brown University Ann Roman, Indiana University School of Medicine
Mauro Bendinelli, University of Pisa Robert Rose, University of Rochester
Thomas Bleck, Northwestern University Kathy Rundell, Northwestern University
Allan Brasier, University of Texas Medical Branch Christian Sauder, United States Food and Drug
Thomas Chambers, Merck & Co. Administration
James Champoux, University of Washington Brad Sefton, University of California at San Diego
Richard Condit, University of Florida Peter Shank, Brown University
Kevin Coombs, University of Manitoba Guido Silvestri, University of Pennsylvania
James Crowe, Vanderbilt University Steven Specter, University of South Florida
Daniel DiMaio, Yale University Mario Stevenson, University of Massachusetts
Ellie Ehrenfeld, National Institutes of Health Medical School
Malcolm Fraser, University of Notre Dame Sankar Swaminathan, University of Florida
Adolfo Garca-Sastre, Mount Sinai School of Peter Tattersall, Yale University
Medicine Gary Tegtmeier, University of Kansas at Kansas City
Rebecca Horvat, University of Kansas at Kansas City Paul Wanda, Southern Illinois University at
Lou Laimins, Northwestern University Edwardsville
Duncan McGeoch, University of Glasgow Scott Weaver, University of Texas Medical Branch
A. Dusty Miller, University of Washington Raymond Welsh, University of Massachusetts
Medical School
Edward Niles, University at Buffalo
David Ornelles, Wake Forest University

Baltimore, D. 1971. Expression of animal virus genomes. Bacteriol. Rev. 35:235241.
This page intentionally left blank
1 A Selective History on the Nature of Viruses
2 Biosynthesis of Viruses: an Introduction to
Virus Classification
3 Modes of Virus Infection and Disease
4 Host Defenses and Viral Countermeasures
This page intentionally left blank

A Selective History on the

Phage Growth: Eclipse and Replication
Nature of Viruses

More than 90% of all human illnesses may be caused by virus infections.
To realize the validity of that estimate, we need only think of our common
colds, influenza, various infirmities due to herpesviruses (e.g., genital her-
pes, infectious mononucleosis, chicken pox, and shingles), virus pneumo-
nias, hepatitis, acquired immunodeficiency syndrome (AIDS), and some
cancers as well. Indeed, virtually every part of our bodies is susceptible to
virus infection, from our heads (viral meningitis and encephalitis) to the
soles of our feet (plantar warts). Until relatively recently, we would also
have had to number among agents of our common infections the viruses
responsible for measles, mumps, rubella, and poliomyelitis, and earlier
still, smallpox. In those instances, we can be grateful that we now have ef-
fective vaccines.
Even if the above estimate of 90% still seems high, the frequency of
viral infection is actually much greater still, since most infections by vi-
ruses are either very mild or not even symptomatic and therefore go un-
noticed. This is an important fact that is discussed in later chapters. Nev-
ertheless, it was the frequency and severity of viral diseases in humans, as
well as in domestic animals and plants, which led to early interest in virol-
ogy. Indeed, were it not for the diseases caused by viruses, the discovery of
these agents (which lie beyond the range of the best light microscopes) at
the end of the 19th century may well not have occurred until the age of
molecular biology, more than half a century later. As it happened, how-
ever, the development of the field of molecular biology itself owes much
to the earlier discovery of viruses.
The development of the science of virology follows two somewhat sep-
arate tracks. Early studies in virology were driven by medical concerns.
Later, many of the major advances in virology developed in the context of
the new field of molecular biology. Today, these two tracks often intersect,
with advances in the molecular and cellular aspects of virus infection
leading to new insights into viral disease processes and rational approaches
to the prevention and treatment of virus diseases.

At this point the reader may be looking for a definition were resistant to subsequent episodes of this disease and
for viruses. However, many virologists believe that al- that an effective control measure (i.e., variolation) was de-
though one can easily acquire a good understanding of veloped based on that observation. Variolation encom-
the unique nature of viruses, one cannot readily pin down passed risks that would be unacceptable today, in particular
a concise definition for these entities. For that reason, a fatality rate of 1 to 2%. However, those risks were ac-
rather than by beginning with a definition for viruses, we ceptable in 18th century Europe when, for example, as
start by establishing their unique properties. Those prop- many as 1 person in 10 died of smallpox.
erties are revealed in a logical sequence from a consider- A major leap forward in preventing smallpox came
ation of some major findings in the history of the field. with the development of the smallpox vaccine in 1798 by
And because that history, which began with the medical an English country doctor, Edward Jenner. Jenners break-
concerns noted above, is interesting in its own right, we through, though still a half-century before the proposi-
start there. tion of the germ theory of disease and 100years before the
actual discovery of viruses, might be regarded as the most
astonishing achievement in the history of medical virol-
THE EARLY YEARS: DISCOVERERS ogy. Jenner learned that some of his patients, who hap-
AND PIONEERS pened to be milkmaids, were resistant to smallpox. His
Choosing a precise beginning for the history of the sci- great insight was to realize that the milkmaids resistance
ence of virology is somewhat arbitrary, in part because to smallpox somehow came about because they had ear-
several illnesses that now are known to result from virus lier acquired cowpox, which is caused by what is now rec-
infections had been recognized for thousands of years ognized as the cowpox virus that infects cattle. (Cowpox
without any knowledge of viruses. Regardless, there is virus is now known to be a relative of the smallpox virus,
some justification for beginning 1,000 years ago with commonly referred to as variola virus.) This led Jenner to
smallpox, because that was when an empirically based mea- carry out an experiment in which he inoculated a child,
sure to control this dreaded disease first became known. James Phipps, with extract from a cowpox lesion and then
That measure was variolation, whereby uninfected indi- demonstrated that young Phipps was resistant to a subse-
viduals were inoculated with material from the scabs of quent challenge with smallpox (Box1.1).
individuals who survived smallpox infection. It arose in More than a century would have to pass before it could
the Far East and has been carried out for the last millen- be appreciated that this protection depended on the facts
nium in China and India. What might have inspired the that cowpox virus is immunologically cross-reactive with
emergence of this effective prophylactic countermeasure smallpox virus and that it produces a relatively benign in-
against smallpox in 11th century China and India? Remark- fection in humans.
ably, it was the practical application of the even earlier real- The term vaccination is based on the Latin word for
ization by the Chinese that individuals may experience cow (vacca) and was coined by Pasteur in recognition of
some illnesses only once in a lifetime. Variolation began to Jenners contribution. The term vaccine now refers to
be practiced in Europe in the early 18th century, after Lady specific and actively immunizing agents that protect
Mary Wortley Montague, the wife of the British ambassa- against a variety of pathogenic microbes. Vaccines are one
dor to Turkey, had her children undergo variolation. of the most effective and widely used prophylactic mea-
Despite its risks, variolation generally resulted in an in- sures ever developed.
fection that was milder than a naturally acquired one, and Interestingly, vaccinia virus, which is the name for the
the variolated individuals were protected for life against virus used in the current smallpox vaccine, is clearly dif-
the more severe naturally acquired smallpox infection. ferent from the cowpox virus in Jenners original vaccine,
Why might exposing an individual, usually a child, to the and its precise origin is not known. At one time, it was
dried-out scabs on the skin of a patient recovering from thought that the cowpox virus was inadvertently replaced
smallpox protect that child from a natural infection? Today by an attenuated (i.e., weakened; see below) smallpox vi-
we understand that the virus in a dried-out lesion of a re- rus. This might easily have happened since the vaccine
covering person has been partially inactivated by that per- was originally propagated by serially passing it from per-
sons immune response, as well as by the drying itself. Yet son to person, a practice since terminated because it was
that partially inactivated virus can still induce immunity to associated with the transmission of other diseases, includ-
natural infection in an individual undergoing variolation. ing syphilis and hepatitis. However, more recent studies
Some readers might find it remarkable that in presci- show that vaccinia virus, although related to smallpox vi-
entific times, and in a non-Western society as well, it was rus, is different from the smallpox virus, perhaps reflect-
recognized that survivors of smallpox, whatever its cause, ing the now traditional practice of propagating the vaccine
History of the Nature of Viruses 5

Box 1.1
Although history usually credits young James Phipps as known. We can only speculate on how Mrs. Jenner might
the first person to be vaccinated, Phipps in fact was not the have regarded these activities. Also, little is known about
subject of Jenners first experiment. Instead, it was Jenners James Phipps, who was only 8years old when he was the
first son, Edward, Jr., born in 1789, whom Jenner inoculated subject of Jenners later experiments. Additionally, nothing
with swinepox virus when the infant was only 10months is known about Jamess parents and whether they may have
old. Jenner of course did not know about microbes, and he consented to Jenners use of James. But Jenner looked after
left no records that might have revealed his purpose in in- James in later years and may have felt some remorse, since
oculating Edward Jr. with swinepox. It may be relevant that he eventually built a cottage for James and even planted
cowpox was relatively rare, while a similar pox disease was flowers in front of it himself. Also not often mentioned,
more common in pigs. Regardless, the baby became sick on Jenner used several other young children in his experiments,
the eighth day with a pox disease, from which he recovered. including his 11-month-old second son, Robert, in addition
At several unspecified times afterwards, Jenner in point to Edward Jr. and James. One of these children died from a
of fact tried to infect Edward Jr. with actual smallpox. He fever, possibly resulting from a contaminant (streptococ-
failed each time, most likely because the swinepox virus may cus?) in the vaccine. Thankfully, such experiments cannot be
indeed have immunized Edward Jr. against smallpox. done today. Yet because of Jenners efforts, this once dread
human scourge now exists only in the laboratory.
Admittedly, this episode involving Edward Jr. somewhat
muddies the waters regarding the otherwise straightfor- This sidebar gives only a glimpse into the fascinating story
ward history of the discovery of smallpox vaccination that behind Jenners work. For more details, I recommend
is usually recounted. Since Jenner left no records of his Virus Hunters by Greer Williams, Alfred A. Knopf,
experiments using his infant son, few details are actually New York, 1960.

strain on the skin of horses. That is, the modern smallpox

vaccine strain may be a horse virus!
Box 1.2
In 1885, still prior to the recognition of viruses as dis-
When contemporary vaccinologists develop vaccines
tinct entities, Louis Pasteur developed a vaccine against
to protect against viral diseases, they are essentially tap-
rabies. Pasteurs vaccine was the second human vaccine ping into biological mechanisms that already have been
and the first deliberately attenuated one. Attenuation is the perfected through eons of natural selection. With many
conversion of a killer microbe into something that is less different viruses, natural infection regularly results in
able to cause disease, yet still able to induce protection. In lifelong immunity against the same virus. This is the
this instance, attenuation was achieved by serial passage of principal fact exploited by vaccinologists. With that in
the rabies agent in rabbits, followed by aging it invitro mind, note again that vaccines against smallpox and
(see Chapter 10). Pasteur had no way of knowing what the rabies were developed years before viruses were even
underlying basis for attenuation actually was, nor did he discovered. Also, in Pasteurs day, nothing was known
even know the viral nature of the rabies agent, since the about how the human body generates immunity. It may
existence of viruses as a distinct class of microorganisms be somewhat disquieting to know that even today, when
remained to be discovered. Knowledge of genes and mu- so much more is known about viruses and our immune
tations, like viruses, also was still in the future. Surpris- response, the development of vaccines remains largely
ingly, even today virologists know relatively little about empirical.
the determinants of virus virulence (Boxes 1.2 and 1.3).
At this point in our history we go back to the mid-19th
century and consider some important developments lead-
ing to the acceptance and success of the germ theory of that different microorganisms are associated with differ-
disease. By this time, the existence of a variety of microor- ent kinds of fermentations (Box1.3). At about the same
ganisms, including bacteria, fungi, and protozoa, was well time, Joseph Lister, an English surgeon who admired Pas-
established. In 1867 Pasteur proposed that microorgan- teurs work on fermentation and spontaneous generation,
isms might produce different kinds of diseases. This reasoned from Pasteurs demonstration of the ubiquity of
premise was based on his earlier experimental findings airborne microorganisms that infections of open wounds

Box 1.3
It is interesting that Pasteur, who was one of the greatest rabies agent. But he later found that he could isolate the
of all experimentalists, did not advance our understanding same microbe from the saliva of healthy children. What
of the nature of the rabies agent. That is, he never came to was this microorganism seen by Pasteur? Ironically, it was
realize that he was dealing with a new category of microbes, Pneumococcus pneumoniae, a major bacterial pathogen that
ones that are submicroscopic and cannot be cultivated on was correctly identified several years later by Albert Frankel.
nutrient media. This is all the more surprising since at the Thus, Pasteur missed the opportunity to identify a human
time the Pasteur Institute was established, rabies actually pathogen much more important than rabies virus, as well as
was its main interest. Thus, Pasteur indeed attempted to find to identify a whole new class of pathogens: viruses. Pasteur in
the rabies microbe, presuming that once found it could then fact was frustrated by his rabies studies, largely because the ra-
be cultivated invitro and serially passaged from one experi- bies agent passed through filters that previously had retained
mental animal to another. He was wrong on the first count, bacteria (see the text). Yet he did correctly surmise that rabies
but right on the second. Unfortunately, he got sidetracked is caused by an unusually small microorganism, beyond the
along the way. In 1880, he injected a rabbit with the saliva of range of his microscopes. Still, Pasteur did not suspect that it
a child who had died of rabies, and then, he examined the was in any basic way different from the pathogenic bacteria
blood of the rabbit after it too died. Under his microscope, isolated previously. Rabies virus was identified in 1903 by
Pasteur in fact observed a microbe in the rabbits blood. It Paul Remlinger, who correctly demonstrated its filterability
had a mucous capsule, and Pasteur thought it might be the at the Constantinople Imperial Bacteriology Institute.

are due to microorganisms in the environment. The aseptic so-called Chamberland filters. These filters, made of un-
techniques that Lister then introduced dramatically re- glazed porcelain, contain pores that are too small to per-
duced infections during surgery. Lister also contributed mit the passage of most bacteria. (The use of porcelain
the technique of limiting dilution to obtain pure cultures filters to retain and concentrate microbes became stan-
of bacteria, while Robert Koch developed the isolation of dard practice after 1884. Chamberland is said to have devel-
bacterial colonies on solid culture media. In addition, oped these bacterium-proof filters while experimenting
Kochs postulates provided the experimental basis for es- with a broken clay pipe purchased from his tobacconist.)
tablishing that a specific microorganism is responsible for It is important to appreciate that Ivanovskys research
a specific disease. Koch had been studying anthrax, a dis- actually might have provided an experimental precedent
ease of cattle, caused by Bacillus anthracis, which can be for demonstrating the presence and activity of a new type
transmitted to people. He used his isolation technique to of infectious agent, namely, viruses. However, as the story
establish pure cultures of a single species of bacteria from goes, Ivanovsky, like Mayer before him, was unable to
the infected cattle. Then, he injected a sample of the pure grow the organism from the infected sap. Thus, he was
culture into healthy animals. The healthy animals then de- unable to satisfy an important component of Kochs pos-
veloped anthrax. Finally, he reisolated the infectious agent tulates (that is, the cultivation of a single species of micro-
from the inoculated animals. This sequence of isolation, organism in pure culture). The influence of the Koch
infection, and reisolation constitutes Kochs postulates. paradigm was so strong that Ivanovsky did not want to
We now come to those developments that led to the consider that he might actually have seen evidence for a
realization that there exists a class of microbes that are previously unknown kind of microorganism. Instead, he
fundamentally distinct from the previously recognized questioned the reliability of his experimental procedures.
bacteria, fungi, and protozoa. Beginning in 1887, a Rus- Perhaps the causative agent was a bacterium and the fil-
sian scientist, Dmitry Ivanovsky, was looking into the ters were defective, or perhaps the causative agent was a
cause of tobacco mosaic disease, so named because of the toxin, a nonreproducing poisonous substance produced
dark and light spots on diseased tobacco leaves. Repeating by an organism. Either of these possibilities could readily
the work of an earlier scientist, the German Adolf Mayer, have been tested. Do you see how?
Ivanovsky showed that the sap of diseased plants con- An important leap forward came in 1898 from the
tained an agent that could transmit the disease to healthy Dutch microbiologist Martinus Beijerinck, who was work-
plants. But Ivanovsky went an important step further. He ing with Mayer but was unaware of Ivanovskys findings.
found that the infectious agent could actually pass through Beijerinck, like Ivanovsky, found that the sap from diseased
History of the Nature of Viruses 7

tobacco plants retained its ability to transmit the tobacco

mosaic disease after filtration through Chamberland fil-
Box 1.4
ters. But Beijerinck went yet another major step further.
I find the genealogy of leading scientists to be interest-
He demonstrated that the sap did not lose its ability to
ing, since influential scientists frequently were trained
cause disease upon being diluted from repeated inocula- by mentors who likewise made extraordinary contribu-
tions or passages through new, healthy plants. This exper- tions. For a compelling example of this phenomenon,
imental result implied that the disease-causing agent was note the background of Max Delbrck, as well as the
in fact replicating in the plant tissue, thus accounting for accomplishments of some who came under his influ-
its ability to replenish its pathogenic activity. This finding ence, as recounted in the text. However, in this sidebar, I
provided the means for distinguishing this new type of want to highlight that Friedrich Loeffler was trained by
pathogenic agent from nonreproducing toxins, which al- Robert Koch. In 1884, 14years before Loeffler isolated
ready were known to be produced by certain bacteria. the foot-and-mouth disease virus, Loeffler used Kochs
So now we have established two fundamental proper- postulates to identify the bacterium that causes diphthe-
ties that are characteristic of this new class of pathogens. ria, Corynebacterium diphtheriae. In addition, Loeffler
First, they are smaller than bacteria, since they pass made the important observation that when he injected
through filters that block bacteria. Second, they require C. diphtheriae into animals, the microbe did not need to
living cells or tissue to support their propagation. The first spread to the tissues that it damaged. This observation
of these properties, their small size, accounts for the fact led Loeffler to propose that the bacteria were producing
that these agents were not seen by the microscopy proce- a poison or toxin that spread to the remote sites. This
dures of the day, which readily revealed bacteria. The sec- was a new concept and a prediction that was later con-
ond property accounts for the inability of Ivanovsky to firmed. Interestingly, and as explained in Chapter 2 and
propagate the tobacco mosaic virus (TMV) in pure cul- in Box1.9, a virus indeed plays a role in diphtheria, in a
ture, that is, in the absence of susceptible living host tis- rather unexpected way.
sue. Still, this part of the story is not yet complete, since it
was not yet clear whether these newly discovered agents
might be liquid or particulate. Indeed, during the first de-
cades of the 20th century they were referred to simply as virus, which was the first known tumor virus, as well as
filterable agents. The issue was not settled until 1938, the first retrovirus to be discovered. More is said about
when the first electron micrographs of TMV were taken that in Chapter 20. For now, see Box1.5. Fourth was the
and the active agents were revealed to be tiny particles. discovery of bacteriophages in 1915 by Frederick W. Twort
The term virus, from the Latin word for poison, came to in London, and independently in 1917 by Felix dHerelle
be used to refer to the agents having the properties de- in Paris. As has happened in other instances, there were
scribed by Mayer, Ivanovsky, and Beijerinck. quarrels over who made the discovery of bacteriophages
Returning briefly to the earlier pioneering days, we first. Regardless, it was dHerelle who gave the bacte
note some other significant achievements. First, in 1898 riophages (phages for short) their name, which actually
Loeffler and Frosch isolated the first virus obtained from means bacteria eaters, which, of course, is not what actu-
animals, the foot-and-mouth disease virus (Box1.4). Sec- ally happens. Nevertheless, since bacteriophages indeed
ond, in 1901, in Cuba, the U.S. Army doctor Walter Reed may kill their bacterial host cells, dHerelle spent a num-
isolated the first virus pathogenic in humans, yellow fever ber of years trying to develop them for use as therapeutic
virus. That virus provided a new surprise. Rather than be- agents in the treatment of bacterial disease. Those efforts
ing transmitted directly from person to person, it was never bore fruit (Box1.6).
transmitted by mosquitoes, a hypothesis advanced earlier
by Cuban physician Carlos Juan Finlay and subsequently
confirmed by Reed. (The affirmation that yellow fever is THE FIRST STIRRINGS OF THE
transmitted by mosquitoes provides one of the more fas- MOLECULAR ERA
cinating stories in the history of infectious diseases. It is An experimental finding that provided a major impetus
told more completely in Chapter 7.) Third, and yet an- towards the eventual emergence of the modern science of
other surprise, was the discovery by Peyton Rous in 1911, virology, and indeed of molecular biology as well, in-
at the Rockefeller Institute (now University) in New York, volved bacteria rather than viruses. The experiment was
that viruses can cause cancer. Rous found that sarcomas carried out by Frederick Griffith in 1928 and involved
(cancers of connective tissue) in chickens could be trans- the pneumococcus bacterium Diplococcus pneumoniae.
mitted by a virus that is now known as the Rous sarcoma Wild-type or smooth D. pneumoniae organisms contain

Box 1.5
It is noteworthy that much of what is known today about was discovered 3years earlier in 1908 by Vilhelm Ellereman
the cellular and molecular nature of cancer, as well as of the and Olaf Bang. They found that leukemia in birds could
mechanisms that control normal cell growth and prolifera- be transmitted by a filterable agent from leukemic cells or
tion, came from studies involving tumor viruses. These by serum from leukemic birds. However, leukemia was not
issues are described more extensively in Chapters 15 and 20. then recognized as cancer, so the significance of this discov-
For now, it is noteworthy that Rous was awarded the Nobel ery went unrecognized. Sarcomas certainly were recognized
Prize for his 1911 discovery, but not until 1966! The 55-year as cancers, but, bearing in mind the level of knowledge at
lag between Rouss achievement and his recognition by the the time, Rouss discovery was seen as little more than an
Nobel Committee is the longest on record. Nobel prizes are interesting oddity. Thus, given the level of understanding
not awarded posthumously. Fortunately, Rous had longevity during the first half of the 20th century, Rouss discovery
on his side. He died 4years after receiving the Prize, at age 87. could not have led to any breakthroughs; certainly not in
1911. Rous abandoned work on the Rous sarcoma virus
Why do you suppose Rous had to wait so long for his con- around 1915 (see Chapter 20), but 20years later he contrib-
tribution to be recognized by the Nobel Committee? The uted significantly to the realization that the Shope papil-
following might contain the seed of an answer. An onco- lomavirus (the first known DNA tumor virus) causes cancer
genic virus closely related to Rous sarcoma virus actually in rabbits (see Chapter 16).

Box 1.6 live virulent pathogens. The following two possible expla-
nations were considered. (Are there any others?) First,
The novel Arrowsmith, by Sinclair Lewis, describes
some substance may have passed from the dead virulent
fictional efforts to develop bacteriophages for use as anti- organisms to the live avirulent ones, enabling the latter to
bacterial therapeutic agents. I highly recommend it, if for produce the polysaccharide capsule. Alternatively, some
no other reason than to get a sense of what it was like to substance may have passed from the live avirulent organ-
do medical research in that earlier time. Interestingly, the isms to the dead virulent ones, restoring the viability of
idea for Arrowsmith was suggested to Sinclair Lewis by the latter. The first explanation was correct, as later dem-
his friend Paul de Kruif, then a scientist at the Rockefeller onstrated in an experiment in which a cell-free extract
Institute, and later a well-known science writer. Also, as prepared from the dead virulent cells was able to trans-
noted in Chapter 2, there is currently renewed interest in form some live avirulent cells to virulence. The reverse
developing bacteriophage-based approaches for antibac- transfer did not yield any viable organisms. (What might
terial therapy. Griffiths purpose have been in carrying out so bizarre an
experiment? It was, in fact, part of Griffiths efforts to make
a vaccine to prevent the pneumonia epidemics that were
occurring at the time. Toward that end, he inoculated mice
with live avirulent and with heat-killed virulent samples of
polysaccharide capsules that enable them to avoid phagocy- D. pneumoniae, as well as with both concurrently. Although
tosis by host macrophages. This capsule is thus a virulence Griffiths experiment is sometimes recognized as one of
factor that enables the bacteria to infect the host and cause the keystones of modern molecular biology, he died in
disease. Nonencapsulated or rough mutants are nonpatho- 1941, never knowing the impact it eventually would have.)
genic. In Griffiths rather bizarre experiment, he inoculated The next part of the story concerns the identification
mice with a mixture of live nonpathogenic mutant bacteria of the crucial transforming factor. One could argue that it
and heat-killed wild-type bacteria. Neither the nonpatho- is genetic in nature, since the transformed bacteria yielded
genic mutant bacteria nor the killed wild-type bacteria progeny cells with the transformed phenotype. Moreover,
alone induced disease. Remarkably, mice inoculated with the transformed bacteria and their progeny behaved
the mixture not only developed severe septicemia but also like the original wild-type bacteria in subsequent trans-
released live virulent encapsulated pathogens. Thus, some forming experiments. In 1944, Oswald Avery, Colin Mac
interaction occurred in the mouse between the live aviru- Leod, and Maclyn McCarty at the Rockefeller Institute
lent bacteria and the dead virulent organisms to generate carried out an experiment based on the earlier finding
History of the Nature of Viruses 9

that transformation is possible using cell-free extracts.

They prepared a whole-cell extract from the encapsulated
Box 1.7
cells and fractionated it into its various macromolecular
The heroic effort that went into purifying and crystalliz-
species, generating protein, lipid, polysaccharide, and de-
ing TMV cannot be appreciated from the brief account
oxyribonucleic acid (DNA) fractions. Next, they asked in the text. However, the Nobel Committee did appreci-
which of these fractions might have transforming activity. ate it, making Stanley, in 1946, the first virologist to be
To the surprise of almost everyone, only the DNA fraction awarded the Nobel Prize (he shared the Prize in Chemis-
was capable of transforming nonencapsulated cells into try). However, even in this instance, it was impossible for
capsulated ones. many scientists of the day to accept that a crystal could
It is important to our story to note that these remark- have the capacity for life. In the minds of some, the specter
able findings, and the conclusion they implied, were met of possible contamination hung over this work. Also, many
with widespread skepticism. To understand why, one must physician scientists of the day did not care about tobacco
adopt the mind-set of the time. First, it was difficult for plants with mottled skin, nor did they appreciate its
the classically trained Mendelian geneticists of the day to possible relevancy to disease in humans.
accept the validity of these strange new experiments. The
experiments of the classical geneticists, which involved Stanley continued to persevere, but on one important
the breeding of generations of plants or animals, estab- account he was wrong. Rather than being comprised
lished that hereditary information is carried in units called entirely of protein, as Stanley had believed, TMV was
genes, and these scientists thought in terms of genes, not shown by Frederick Bawden and Norman Pirie to also
in terms of molecules of nucleic acid. More importantly, contain RNA. The composition actually was 94% protein
in those days, DNA was looked upon as a structurally un- and 6% RNA, moving TMV somewhat away from being
interesting, monotonous molecule, rather like a starch. In merely a chemical to somewhat closer to being an organ-
contrast, the wide variety of proteins seemed to provide a ism. Interestingly, Max Schlesinger in 1933 was actually
virtually unlimited number of possible genes. Thus, it was the first person to find a nucleic acid in a virus. Making
use of high-speed centrifuges, he purified a bacteriophage
widely assumed that the genetic composition of chromo-
to high purity and demonstrated that it is comprised of
somes depended on their proteins, not their nucleic acids.
50% protein and 50% DNA. However, Schlesinger did
Consequently, the classical geneticists did not see a con-
not crystallize his virus, as Stanley had done, and his
nection between the biological unit, the gene, and the
discovery attracted much less attention than that of
chemical unit, the DNA. Thus, despite these experiments, Bawden and Pirie.
the notion that protein constituted the genetic material
generally persisted.
One objection raised against the conclusion that the
transforming activity resided in DNA was that a minute
amount of protein, undetectable by the methods em-
ployed, might have remained associated with the DNA the most part medical and agricultural. Essentially all that
during fractionation. This protein might then have been was known about the biology of viruses was that viruses
responsible for transforming activity. However, Avery, are smaller than bacteria and that they can propagate only
MacLeod, and McCarty showed that extensive protease withinsuitable host cells. Moreover, the study of viruses
treatment of the extract had no effect on its transforming had not yet advanced biological knowledge in general.
activity, whereas even very brief exposure to nuclease However, biochemists had been making great strides in
completely destroyed biological activity. Another objec- purifying and crystallizing proteins. (Note that solving the
tion was that even if DNA was responsible for transfor- structure of proteins by crystallography was still well be-
mation in this instance, perhaps it acted in some non yond the technology of the day.) Inspired by the success of
genetic chemical manner to affect capsule formation. This the protein crystallographers and armed with indirect
objection was addressed by the experiments of Rollin evidence that TMV is at least partly a protein, Wendell
Hotchkiss, who demonstrated that transformation also Stanley at the Rockefeller Institute succeeded in crystalliz-
works for bacterial characteristics (e.g., antibiotic resis- ing TMV. Two important points now need to be noted.
tance) that have nothing to do with capsule formation. First, the protein crystals of TMV indeed were endowed
At this point, we take a short step back in time to 1935 with the infectious activity of TMV. Second, crystals are
to consider a major achievement in virology that would exquisitely pure. Thus, it could now be unequivocally
dramatically influence later developments. First, we remind demonstrated by others (Box1.7) that TMV is not a pure
ourselves that up to this time, interest in virology was for protein but instead contains about 6% ribonucleic acid

(RNA) (see below). Consequently, whatever it was about Max Delbrck was a key player in this new group of sci-
TMV that enabled it to produce copies of itself, that abil- entists, and he is recognized as one of the principal found-
ity resided in its protein, or in its nucleic acid, or in a com- ers of the new science of molecular biology. Delbrck orig-
bination of its two constituents. (Also note that the ability inally trained as a physicist in Germany during the 1920s,
of TMV to form crystals implied that it had a regular studying quantum mechanics under the guidance of Max
structure. This is discussed more fully later.) Born. Moreover, he interacted with many others of the great
Arguably, Stanleys achievement was a major milestone physicists of the day, including Wolfgang Pauli, Albert
in the history of not only virology but also biology. If vi- Einstein, and Erwin Schrdinger. In 1931 Delbrck went to
ruses are so simple that they could be crystallized like table Copenhagen for postdoctoral studies with the great Niels
salt and yet express that most fundamental property of Bohr, and it was actually Bohr who aroused Delbrcks in-
living systems, the ability to replicate, then there was rea- terest in biology. After a short subsequent stay in England,
son to believe that the nature of biological replication Delbrck returned to Germany in 1932 to work as an as-
might eventually be understood in chemical terms. To ap- sistant to Lise Meitner, who eventually discovered nuclear
preciate how this concept stirred the souls of some scien- fission (Box1.8). But while he was with Meitner, Delbrcks
tists of the day, we once again adopt the mind-set of the focus was actually on developing quantum mechanical
times. Whatever the chemical nature of the genetic mate- models of genes, his thinking about the nature of genes ac-
rial might be, it would have to contain information and tually having been influenced by Schrdinger. By 1937, the
also have the ability to be precisely copied. Because of the situation in Nazi Germany became intolerable for Del-
general ignorance regarding the structures of large mole- brck, and it was then that he began the journey that would
cules in the cell and the then-prevailing belief that the ge- lead to his move to the California Institute of Technology
netic material was protein, it was felt by many that it (Caltech), where he initiated developments of extraordi-
would be essentially impossible to understand heredity in nary importance for virology, which as noted above, led to
chemical terms. Attempts were made to account for how a the emergence of the new science of molecular biology.
protein might replicate, but as you might presume, these Delbrck began to work with bacteriophages in 1938,
attempts did not lead to especially satisfying models. Thus, recognizing that these viruses would be an ideal system
many serious biologists and chemists expressed the belief for probing the nature and action of genes. Indeed, mod-
that some vital force outside the known laws of chemis- ern phage research began when Delbrck entered the field.
try would be needed to explain the phenomenon of life. This was the result of Delbrcks own experiments, and
This doctrine was referred to as vitalism, and it still had also the work of other members of the phage group that
its serious adherents up to the time of Stanleys work. Delbrck and the Italian scientist Salvador Luria brought
However, crystallography is a very precise science, and the together at Caltech.
fact that TMV could be crystallized and yet retain biological Delbrck actually began his foray into genetics by fol-
activity strongly implied to at least some scientists that lowing up on H. J. Mullers discovery in the 1930s that
physical and chemical explanations indeed would suffice X rays and ultraviolet (UV) irradiation could cause muta-
to explain life. Thus, Stanleys achievement would eventu- tions in fruit flies. To account for Mullers findings, Del-
ally mark the death knell of vitalism and spur the begin- brck spent several years working towards a quantum
ning of the field of molecular biology (Box1.7). physics interpretation of gene mutation, which ultimately
led to his thinking of genes as molecules, rather than as
abstract entities, a first step on the path to molecular biol-
THE PHAGE GROUP ogy. Luria, in turn, was excited by the new idea of genes as
Spurred on by this new line of thought, a somewhat atyp- molecules and sought a system in which to pursue their
ical group of investigators sought to understand the nature analysis. Independently of Delbrck, he too thought that
of genes. They were atypical in that many had little or no bacteriophages might be an ideal system to probe the na-
knowledge of traditional genetics, or biochemistry, or even ture of genes, since bacteriophages grew rapidly and the
biology in general. Many were physicists by background, effects of radiation on them could be measured with pre-
but they had a single goal in mind: to understand the phys- cision. Luria had hoped to be able to work with Delbrck.
ical basis of the gene. Important to our story, several recog- In the late 1930s, Luria, who was Jewish, faced persecution
nized the advantages of focusing their research efforts on in then-fascist Italy. Independent of each other, Luria and
viruses. This odd groups interest in genes and its focus on Delbrck each left Europe and finally met in the United
viruses would lead to discoveries of singular overwhelm- States in 1940. That summer, at the Cold Spring Harbor
ing importance. Indeed, their research approaches and the Laboratory on Long Island, they began their first experi-
results they generated gave rise to molecular biology. ments on the life cycle of bacteriophages.
History of the Nature of Viruses 11

Box 1.8
Shortly after Delbrck left Nazi Germany, Meitner went on German agents in Stockholm were out to assassinate Bohr
to discover nuclear fission; this story has fascinating rami- led to his harrowing escape to England. He was spirited
fications because Meitner, who was Jewish, was by then a away in the unpressurized empty bomb rack of a Royal Air
nonperson in Germany. Meitner was able to find safe haven Force Mosquito Bomber, in which he lost consciousness
in Holland, thanks to the efforts of Dutch physicists who and nearly died from a lack of oxygen. Bohr spent the last
persuaded their government to admit her without a visa, on 2years of the war in England and America, where he was
an Austrian passport that no longer was valid for her. Niels associated with the Atomic Energy Project, and then became
Bohr subsequently found a laboratory for Meitner in one of the first and most prescient arms control advocates.
Sweden where she might continue her work, which was These incidents are chronicled in The Making of the Atomic
financed by a grant from the Nobel Foundation. Bomb by Richard Rhodes (Simon & Schuster, New York,
NY, 1986).
During the Nazi occupation of Denmark in World War II,
Bohr himself was endangered, and so he too escaped to Delbrck (jokingly?) took indirect credit for Meitners
Sweden. From there, Bohr played a crucial part in the rescue discovery of nuclear fission, saying that his waning interest
of nearly all of the Danish Jews by pressuring the Swedish in physics was holding back Meitners group, and thus, his
government to accept responsibility for them. Rumors that leaving enabled the discovery to happen.

The original phage group founded by Delbrck and particle that is present in the sample will infect a cell in the
Luria trained a brilliant second generation of phage work- dense bacterial lawn that develops on the agar surface.
ers and virologists. What distinguished this group from Each infected cell eventually bursts or lyses, releasing
others was their single-minded purpose to understand the progeny viruses that infect the immediately adjacent cells
physical basis of heredity by analyzing phage replication. in the bacterial lawn. After a few such cycles, a plaque that
(The influence of Delbrck, Luria, and the phage group is comprised of a focus of lysed cells will be visible to the
was enormous. To get a feeling for the extraordinary vi- unaided eye (Fig. 1.1). A virus particle that registers as a
brancy and accomplishments of that group and their era, I plaque is referred to as a plaque-forming unit (PFU).
strongly urge you to spend some time with Phage and the
Origins of Molecular Biology: the Centennial Edition, edited
by J. Cairns, G. S. Stent, and J. D. Watson [Cold Spring Har- Figure1.1 A petri plate showing growth of a lawn of Escherichia coli
bor Laboratory Press, Cold Spring Harbor, NY, 2007].) bacteria on which T2 phages have formed plaques. Reprinted with
permission from G. S. Stent, Molecular Biology of Bacterial Viruses
Phage Growth: Eclipse and Replication (W. H. Freeman and Company, San Francisco, CA, 1963).

Delbrcks first major experiments with phage were car-

ried out with E. L. Ellis. One of these experiments, the so-
called one-step growth experiment, characterized the
parameters of bacteriophage replication. The experiment
was especially important for what it revealed about phage
replication. However, it also was noteworthy for having
marked the beginning of quantitative virology.
We begin with a few words about experimental proce-
dures. In the one-step growth experiment, Delbrck and
Ellis needed to titrate or measure the number of infec-
tious phages in their samples. To accomplish this, they
made use of the plaque assay, a method that was devel-
oped earlier by dHerelle, one of the discoverers of bacte-
riophages (see above). In a plaque assay, appropriate dilu-
tions of a virus sample are added to a dense suspension of
susceptible host bacteria, which is then spread over a sur-
face of nutrient agar in a petri dish. Each infectious virus

The one-step growth experiment then went as follows. experimental findings, in particular, the initial 24-min
A virus inoculum was added to a dense suspension of host lag? The answer is as follows. During the 24min following
bacteria, which then was incubated for several minutes to infection, phage replication was occurring within infected
allow the virus to attach to the cells. Next, unadsorbed virus cells, but none of the new phage were yet being released
in the suspension was removed, thereby synchronizing the from the host cells. Regardless of how many new phage
infection among the individual infected cells. Moreover, may have accumulated within each of the infected cells,
this step ensured that any PFU that were measured during each of those cells, if sampled while still intact, could pro-
the experiment originated from the initially infected cells. duce only one plaque on the lawn of indicator cells. Thus,
Unadsorbed virus can be removed by several means. One the plaques produced by the samples taken over the first
way is by low-speed centrifugation, since the bacteria are 24min were not produced by individual phage particles
pelleted at low speeds while the unadsorbed phages remain in the suspension. Instead, those plaques were generated
in the supernatant fraction. The pellet is then washed and from the initially infected bacterial cells. Then, at 24min
resuspended. The resuspended bacteria are also diluted so the infected cells began to burst open or lyse, thereby re-
that essentially all new viruses that are produced will result leasing the several hundred progeny viruses that they each
from a single cycle of infection. The infected bacterial sus- now contained. Each of the released virus particles could
pension is then allowed to incubate, and at various times then register as a PFU in the plaque assay.
samples are removed from it and plated with sensitive The time elapsing between the moment of infection and
bacteria, as in a standard plaque assay. But note that in the the release of progeny virus is referred to as the latent pe-
one-step growth experiment each intact infected cell, as riod. This is certainly a misnomer, since much is going on
well as each free virus particle in the suspension, produces during this time. The time during which individual prog-
a plaque on the lawn of sensitive bacteria. eny viruses are being released from the lysing cells, thereby
Delbrcks major finding was that for the first 24min becoming available for detection by plaque assay, is referred
after infection the number of PFU in the suspension re- to as the rise period. The average number of PFU produced
mained constant (Fig. 1.2). For the next 10 min, the by each infected cell is referred to as the burst size, about
number of PFU rapidly increased several hundredfold, 100 PFU per infected cell in the example shown.
eventually reaching a plateau. How might one explain these The one-step growth experiment set the stage for in-
vestigating the events occurring during the crucial latent
period, that is, the time during which the parental phage
particles are replicating several hundredfold within the
Figure1.2 The one-step growth experiment of phage T4. In this host cells. In 1948, A. H. Doermann began to examine
particular experiment, exponentially growing E. coli cells were infected
events during the latent period by artificially breaking
with an average of one T4 phage per cell. The phage was allowed to
adsorb to the cells for 2min, after which the bacteria were diluted open the infected cells. This was accomplished by incu-
10,000-fold into fresh media. The cells were then incubated, and sam- bating samples from the inoculum in chloroform. Impor-
ples were removed at various times and analyzed by plaque assay on tantly, this approach enabled Doermann to follow the in-
sensitive bacteria. From A. H. Doermann, J. Gen. Physiol. 35:645652,
tracellular accumulation of phage, each particle of which
1952, as adapted by G. S. Stent and R. Calendar, Molecular Genetics,
2nd ed. (W. H. Freeman and Company, San Francisco, CA, 1958), with could now produce a plaque. Doermanns experiment
permission. gave a most surprising and important result. He found
that during the first 10min after bacteriophage infection,
the input virus disappeared (Fig. 1.3). That is, there were
Infective centers per ml in first growth tube

no PFU present in either the cells or the culture fluid dur-
ing this time, despite the fact that several hundred would
eventually emerge from each infected cell that was allowed
to go the full course of infection. The time that elapses
between infection and the intracellular appearance of in-
106 fectious virus particles is referred to as the eclipse period.
Rise (To be more precise, it is the time that elapses between
Latent period period infection and the appearance of an average of one intra
cellular PFU per infected cell.) So now we have a new in-
triguing question to answer, namely, what happens to the
105 parental virus during the eclipse period?
0 10 20 30 40 The experiment that would eventually answer this com-
Time after infection (min) pelling question was inspired by information obtained by
History of the Nature of Viruses 13

Infectious units per productive cell

Total virus

Rise period
10 Latent period
Eclipse Extracellular
period virus

10 20 30
Concentration of Time (min)
productive cells

0.01 accumulation period

Figure1.3 Diagram of relationships between the eclipse period,

the latent period, and the rise period during one-step multiplication
of bacteriophage T2. The conditions are essentially as for Fig. 1.2.
However, the infection was synchronized and unadsorbed virus
was eliminated by the addition of antivirus antiserum, followed
by centrifugation and resuspension of the bacteria. A sample was
then immediately plated for plaque assay to determine the concentra-
0.1 m
tion of infected cells. At subsequent times other samples were taken.
Each of these samples was then divided into two aliquots: one was
shaken with chloroform to break open the bacteria so that total
virus might be measured by plaque assay; the other was freed of Figure1.4 Electron micrograph showing T4 phages attached to
bacteria by centrifugation, and the supernatant was then assayed to E. coli via their slender tails. The sheath portions of the phage tails
measure extracellular virus. The viral titers are compared with the are contracted, as occurs when the phage DNA is injected into the
concentration of infected cells as 1.0. Thus, the curves indicate bacterium (see Chapter 2). Looking carefully, notice that the phages
the average number of PFU per infected cell over time. Reprinted are attached to the bacterial cell wall by so-called tail pins that extend
from B. D. Davis, R. Dulbecco, H. N. Eisen, and H. S. Ginsberg, from base plates at the ends of the sheaths. Also, notice that tubes
Microbiology, 3rd ed. (Harper & Row, Hagerstown, MD, 1980), extend from the phage tails into the bacterial cell. Moreover, strands
with permission. of phage DNA can be seen extending from the tips of the tubes into
the cell, and the phage heads are now empty. From L. D. Simon and
T. F. Anderson, Virology 32:279297, 1967, with permission.

directly visualizing phage particles. Recall that no one had is the topological relationship between the DNA and the
yet actually seen a virus. However, viewing viruses became protein in the phage particles. Anderson provided the an-
possible with the development of the electron microscope swer in 1949, when he found that he could rupture the
in the late 1930s. In 1942 Salvatore Luria and T. F. Ander- phage heads by osmotic shock (i.e., by sudden dilution of
son took the first electron micrographs of the T-even bac- a concentrated phage suspension in high salt into distilled
teriophages then being studied at Caltech. These micro- water), causing release of the DNA (Fig. 1.5). The electron
graphs showed that the phages are tadpole-shaped entities, micrographs of the ruptured phages clearly demonstrated
with distinct heads and tails. In addition, phages appeared that the phage heads are comprised of an outer protein
to attach to their host cells by their slender tails (Fig. 1.4). shell that encloses the phage DNA within it.
Moreover, the parental phages appeared to remain at- So now, here is the conundrum. If a phage is comprised
tached to the outside of the cell for the duration of an in- of this DNA-containing protein shell that appears to re-
fection. Still, this was only one of the clues as to what main outside the cell during infection, and if the nucleic
might be the basis of the eclipse period. Chemical analysis acid is the genetic material as implied by the transforma-
of the T-even bacteriophages, carried out earlier by Schle- tion experiments, how might that DNA participate in the
singer, demonstrated that these viruses are composed of replication of the phage? Or if the protein were the genetic
roughly equal amounts of DNA and protein. A key question material, as still believed by many, how might it do the

the DNA of the phage was radioactively labeled. This was

accomplished by infecting bacteria that were being culti-
vated in the presence of either 35S or 32P. During T2 repli-
cation in the presence of the isotopes, the bacteriophage
incorporated the radioactive 35S or 32P into their protein
or DNA, respectively. Each preparation of labeled phage
was then allowed to attack fresh bacteria in the absence of
any isotopes. After a short period to allow phage to attach
to the cells, unattached phage was removed by low-speed
centrifugation of the cells and washing of the pellet. The
cells then were resuspended and subjected to the shearing
forces of the Waring blender, which previously had been
shown to strip the phage heads from the cells. The experi-
ment yielded two results of immense significance. First,
80% of the 35S-labeled phage protein could be stripped off
the bacteria, whereas essentially all the 32P-labeled phage
DNA remained with the cells. Moreover, the phage DNA
apparently entered into the cells, since it was resistant to
added DNase. Second, despite being subjected to the War-
ing blender, the bacteria produced progeny phage just as
though they had not been blended. Thus, stripping the
phage protein off the infected bacteria did not prevent the
production of a normal crop of progeny phage. That is,
1 m
once the phage DNA has entered the cell, the parental
phage protein has no further function. Hershey and Chase
provided additional evidence that the phage injects its
Figure1.5 Electron micrograph showing the DNA molecule released DNA into the cells. In particular, they demonstrated that
from a T-even phage particle by osmotic shock. Contour measurements the phage could attach to bacterial cell wall fragments in
show that the DNA is a single continuous molecule that is about
suspension and then release its DNA into the medium on
50 m long, about 550times longer than the width of the phage head
that contains it. From A. K. Kleinschmidt, D. Lang, D. Jacherts, and the other side of the wall fragments. As expected, once re-
R. K. Zahn, Biochim. Biophys. Acta 61:857864, 1962, with permission. leased in this way, the phage DNA was sensitive to diges-
tion by DNase.
The singularly important conclusion that was strongly
implied by the Hershey-Chase experiment is that the ge-
same? Or put another way, what constitutes the infection, netic information of the phage is embodied in its DNA.
and what does this have to do with the eclipse period? An- This conclusion is completely consistent with the implica-
derson provided some additional valuable information tion of the transformation experiments conducted a de-
that would be important in getting to the answers. He sus- cade earlier by Avery, MacLeod, and McCarty. The results
pected that the attachment of the phage to the bacteria via of the Hershey-Chase experiment also explain the eclipse
their slender tails should be unstable. He demonstrated period, since the naked phage DNA that initially enters
that this was the case by shearing the phage off the bacte- the host cell is not capable of initiating a plaque. That is,
ria in a Waring kitchen blender. Now the stage was set for no PFU are present in the infected cells until the appear-
another one of the most famous and significant experi- ance of the first progeny infectious phage particles, in
ments in the history of virology and molecular biology as which the DNA is packaged into a protein coat capable of
well, the Hershey-Chase experiment. initiating infection, thereby marking the end of the eclipse
The above experimental findings provided the impetus period.
for Alfred D. Hershey and Martha Chase to carry out their Importantly, for a stage in the life cycle of T2 phage, all
famous experiment in 1952 at Cold Spring Harbor, Long that exists to connect one generation of phage with an-
Island. In this experiment, Hershey and Chase were able other is a molecule of DNA. This is certainly in marked
to follow the separate fates of the protein and DNA com- contrast to the life cycle of cells, which grow larger and
ponents of the T2 phage during infection. To do this, they then divide. Thus, the eclipse period truly sets phages
first prepared phage stocks in which either the protein or apart from cellular life forms in a fundamental way.
History of the Nature of Viruses 15

We still have to describe the events that occur during

the eclipse period, as well as the mechanism of entry, but
Box 1.9
that will come in the next chapter. At this point it suffices
The work of Victor Freeman in 1951 at the University of
to say that the phage DNA will somehow oversee the pro-
Washington, as well as the work of others, led to the real-
duction of the proteins from which new phage particles ization that the deadly toxins produced by Corynebacteri-
can be assembled. um diphtheriae and Clostridium botulinum indeed are the
Shortly before Hershey and Chase carried out their products of lysogenic phages carried by these bacteria.
classic experiment, Lwoff and coworkers showed that A particularly relevant experimental finding was that
some phages do not necessarily replicate in, and lyse, their avirulent strains of these bacteria became virulent when
host cells. Instead, infection by these phages may lead to infected with phages that were induced from virulent
an outcome in which the phage genome is maintained in strains. In addition to carrying genes for toxins, temper-
a stable state in the host cell, from one cell generation to ate phages also may carry genes for antibiotic resistance
the next. Later, it was found that in some instances the (e.g., penicillinase). So then, are diphtheria and botulism
phage genome is maintained by being integrated into the due to phages or to bacteria?
cellular genome. In other instances, it is maintained as a
plasmid (a DNA molecule that can replicate indepen- It also is interesting to consider the origin of the term
dently of the cells chromosomal DNA). Phages capable of lysogeny. It was as follows. Some bacterial strains,
this lifestyle are said to be temperate, and their stable which were not knowingly infected with any virus, would
relationship with the host cell is referred to as lysogeny. occasionally undergo spontaneous lysis. When doing so,
Temperate phages are able to assume the lysogenic state these bacteria would release bacteriophages. Thus, these
because they have evolved mechanisms that enable them to often cryptic bacteriophages were said to be lysogenic, re-
control their replicative functions. That is, they are able to flecting their ability to lyse their host cells. Consequently,
express only a subset of their genes when in the lysogenic it is a bit odd that their temperate relationship with their
state: specifically, those genes that are necessary to initiate host cells is referred to as lysogeny.
and maintain the lysogenic state. However, as demonstrated
by Lwoff, temperate phages may be induced to come out of
the lysogenic state and lyse their host cells (Box1.9).
Some animal viruses too are able to assume a relation- remains in the aqueous phase). When applied to wounded
ship with the host cell that superficially resembles lyso tobacco leaves, the RNA, but not the protein, produced
geny. Chapter 2 contains a more detailed account of the a crop of complete virus particles. Thus, whereas DNA
molecular mechanisms used by bacteriophages to estab- serves as the genetic material for all cellular organisms,
lish and maintain lysogeny. The mechanisms used by ani- RNA serves as the genetic material for many viruses.
mal viruses to initiate and maintain latency are considered
in later chapters.
We have essentially completed this part of our history DEFINING VIRUSES
of virology, but several items of related interest still need Bearing in mind the characteristics of viruses revealed by
to be noted. First, the results of the Hershey-Chase exper- our historical survey, we return to a question posed ear-
iment convinced all but the greatest skeptics that DNA lier: how might we define viruses? The essential properties
indeed is the genetic material. Moreover, this experiment of viruses that were revealed by our historical account in-
was a key impetus that drove Watson and Crick to solve the clude the following. (i) Viruses are smaller than bacteria
structure of DNA, which they indeed achieved in the next and pass through filters that block the passage of bacteria;
several months. Second, we must note the important fact (ii) viruses can replicate only within suitable host cells;
that TMV, as well as a large number of other viruses (in- (iii) viruses may be comprised only of protein and nucleic
cluding such favorites as human immunodeficiency virus, acid; (iv) the nucleic acid component of many viruses is
influenza virus, and Ebola virus), contain RNA rather than RNA; (v) the nucleic acid of the virus contains its genetic
the chemically related molecule DNA. That is, for these information; (vi) during a stage of its replicative cycle
viruses the RNA portion alone serves as the genetic mate- the virus may exist only in its nucleic acid; and (vii) given
rial. This was convincingly demonstrated for the first time the small size of viruses and their simple structure (i.e.,
by an experiment of Heinz Fraenkel-Conrat and Gerhard they can be crystallized) relative to cells, we rightly would
Schramm, in which they separated the RNA and protein expect that they are dependent on their host cells metab-
of TMV (by shaking the particles in phenol and water; olism to provide energy and raw materials needed for
protein partitions into the phenol, and the nucleic acid their replication.

A compendium of virus characteristics is not a defini- understand the nature of viruses than it is to debate
tion per se. Still, various attempts to define viruses have whether or not they are alive. In fact, the question of
emphasized different combinations of these fundamental whether viruses are alive is neither profound nor a matter
characteristics. Although I believe it is more important to of wonderment. Instead, it really comes down to a matter
know the distinctive features of viruses than to define of definitions. For example, although we readily appreci-
them, a definition that works fairly well is the following, ate that cats and dogs and frogs and even bacteria are or-
quoted from an earlier, superb text by S. E. Luria, J. E. Dar- ganisms and alive, it is not easy to state in any concise way
nell, D. Baltimore, and A. Campbell (General Virology, what it is about them that makes them so. Thus, depend-
3rd ed., John Wiley & Sons, New York, 1978): ing on our rather arbitrary definition of alive or organ-
Viruses are entities whose genomes are elements of nucleic
ism, we might, or might not, include viruses. For example,
acid that replicate inside living cells using the cellular syn- by some accounts, a virus is no more an organism or alive
thetic machinery and causing the synthesis of specialized than a chromosome. The reason given is that viruses, like
elements that can transfer the viral genome to other cells. chromosomes, are dependent on a host cell for their rep-
lication. Thus, by the criterion of independence, viruses
This definition emphasizes the facts that viruses pos-
do not qualify as organisms. Yet by that criterion, the chla-
sess genetic elements that make use of resources inside
mydiae, which are obligate intracellular bacterial para-
their host cells to generate their progeny and that the virus
sites, would likewise also not be considered to be alive. So,
particle exists to transfer the virus genome into other cells.
given the somewhat arbitrary criteria (and their interpre-
However, even this definition omits some of the key points
tations) which underlie our definitions for organisms and
enumerated above, such as the small size of viruses, the
viruses, it is much more important to appreciate what vi-
fact that viruses may have RNA genomes, and the unique
ruses are than to try to definitively answer the question of
eclipse period (Box1.10).
whether they are alive.
For me, the wonderment of viruses comes from con-
sidering the many different adaptations that these entities
ARE VIRUSES ALIVE? have evolved, despite having only a handful of genes in
A favorite question of students is whether viruses are alive. many cases, to solve the numerous problems they face
I feel the same about this question as I do about attempts in successfully exploiting their hosts. Indeed, this point
to define viruses. That is, it is much more important to helps me to personally answer the question we started with.

Box 1.10
Didier Raoult and colleagues recently identified a previously Indeed, it is larger than the genomes of at least 25 known
unknown virus that challenges some of our notions of what bacteria! Consistent with its huge genome, particles of this
a virus ought to be. Their discovery resulted from efforts to virus contain some 438 different proteins. This is equivalent
identify intracellular bacteria that might be associated with to the number of proteins associated with some smaller bac-
amoebae. In this regard, some important human bacterial teria, such as Rickettsia. Moreover, this virus contains genes
pathogens, including Legionella, Chlamydia, Mycobacterium, that were not previously seen in viruses. These genes include
Shigella, Rickettsia, and Listeria, have all adapted an intracel- four that encode transfer RNA (tRNA) synthetases, five in-
lular lifestyle. Strangely, one amoeba-associated putative volved in DNA repair, and three chaperones. Yet the organ-
bacterium, initially classified as belonging to the Legionella- ism indeed is a virus. It lacks ribosomal protein-encoding
like amoebic pathogens, resisted all efforts to amplify its 16S genes and does not appear to contain any enzymes for
ribosomal RNA (rRNA). Believing that this failure was due energy metabolism. Most importantly, it does not undergo
to an inability to digest the organisms cell wall, Raoult binary fission but instead demonstrates typical eclipse phase
and coworkers examined it by electron microscopy, doing viral replication. It was named mimivirus, which is short
so before and after carrying out procedures that would for mimicking microbe. See Box19.4 for an interesting
have extracted the putative cell wall. To their surprise, update on mimivirus.
what they saw was a giant spiked icosahedral virus-like
particle (see Chapter 2). Next, they found that it contained Reference
Raoult, D., S. Audic, C. Robert, C. Abergel, P. Renesto, H. Ogata,
a double-stranded DNA genome of 1,184megabases, which B. La Scola, M. Suzan, and J. M. Claverie. 2004. The 1.2-megabase
is huge in comparison to that of any other known virus. genome sequence of Mimivirus. Science 306:13441350.
History of the Nature of Viruses 17

As Luria, Darnell, Baltimore, and Campbell (1978) note RNA viruses, of dissimilar genetic sequences, contain an
in their book referenced above, a definition for an organ- RNA polymerase gene that is highly conserved among
ism might emphasize its historical continuity, rather than them. Since varieties of these viruses are found in plants,
its functional independence. For example, An organism is as well as across the animal kingdom, they may have
the unit of a continuous lineage with a continuous evolu- evolved in their respective hosts from an ancestral virus
tionary history. In that sense, viruses are most certainly that was present before the separation of kingdoms. Alter-
organisms and alive. A chromosome, by contrast, is a con- natively, the ancestral virus might have evolved to spread
stant component of a cell. It evolves in the context of the to new hosts. We do not know.
cell, which in turn may be part of a multicellular organism
and the actual unit of evolution in that instance. (But as
we soon will see, viruses and their hosts evolve together. THE MODERN ERA OF ANIMAL
For instance, in the case of animal viruses, there is a con- VIROLOGY
tinuing evolutionary dance between the immune defenses Recapping some of the history discussed above, the dis-
of the host and the ability of the virus to evolve counter- coveries of filterable infectious agents beginning at the
measures to those defenses [Chapter 4]. Does that notion end of the 19th century and continuing for the next sev-
support or detract from the view that viruses are alive?) eral decades were motivated by the desire to identify the
etiologic agents of particular infectious diseases in hu-
mans, domestic animals, and plants. In contrast, progress
ORIGIN OF VIRUSES towards actually understanding the nature of viruses in-
Unfortunately, we cannot know the evolutionary history volved mainly studies of bacteriophages and was driven
of viruses, since viruses leave no fossils. Yet viruses may primarily by the desire of the early phage workers to un-
have existed since the origin of cells. That assertion is con- derstand the physical basis for heredity.
sistent with the fact that viruses seem to infect all classes Why then was there little concurrent advancement to-
of cellular organisms, from the simplest bacteria to the wards understanding the biology of viruses that were of
most complex metazoans. medical, veterinary, and agricultural importance? Prog-
We may hypothesize about the origin of viruses, pro- ress towards understanding phages was possible largely
vided we remain aware that these are just guesses. With because of the development of the plaque assay, which en-
that proviso in mind, a likely possibility is that at least abled phage researchers to carry out quantitative studies
some viruses arose from the jumping genes and replicat- in a simple and easy-to-manipulate system. In contrast,
ing extrachromosomal nucleic acids that exist in cells. progress towards understanding the biology of animal vi-
Those entities may have become viruses upon acquiring ruses during those early years was hampered by the fact
the ability to generate coats that would facilitate their cell- that there was no comparable simple experimental system
to-cell transfer and enable their extracellular survival. At in which to study or quantify them. Although some ani-
that point they would embark on their independent evo- mal viruses could be grown in embryonated eggs (e.g.,
lutionary courses. Such a sequence of events might have influenza virus), the major experimental system of the
occurred independently on multiple occasions. time was the whole animal.
Another possibility is that some viruses may have Several major breakthroughs, beginning in the late
arisen from intracellular cellular parasites, which pro- 1940s, dramatically changed this state of affairs. First,
gressively acquired the characteristics of viruses as they John Enders demonstrated that poliovirus could be grown
became more dependent on their host cells. Regarding to a high titer in explanted animal cells. Then, in 1952,
this possibility, see the discussion of the mimivirus in Renato Dulbecco went a major step further when he de-
Box1.10. veloped a plaque assay using cultured cells, which he used
The multiple plausible possibilities suggest that some to accurately quantify Western equine encephalitis virus
virus families have no ancestors that they share with other and poliovirus (Fig. 1.6; Box1.11).
virus families. That is, entities that we now recognize as The plaque assay technique for animal viruses is simi-
viruses may have arisen independently over the years. Ap- lar in principle to the procedure for bacteriophages. How-
ropos that, when we compare the patterns of gene arrange- ever, whereas bacterial cells grow on top of, or within,
ment and expression of the different families of viruses, we agar-containing media, animal cells in culture generally
will appreciate much better the great differences among must attach to a solid surface in order to grow. When ad-
viruses. On the other hand, we should note that even hering to the bottom of the petri dish, under appropriate
seemingly unrelated viruses might show signs of a common nutrient media, animal cells can form a continuous mono-
ancestor. For example, a variety of so-called plus-strand layer of cells. With those points in mind, the plaque assay

only by viable cells. These developments brought about

the emergence of quantitative animal virology.
Importantly, and not mentioned before, all the viruses
in a plaque are descended from a single virus. This fact
enabled the isolation of virus mutants, thereby facilitating
the genetic analysis of animal viruses, as well as phages.
Moreover, it enabled the isolation of avirulent virus mu-
tants for use as vaccines.
Important related developments included the establish-
ment of continuous cell cultures by Earle in 1948 and the
Figure1.6 Plaques produced by Western equine encephalitis virus on development by Eagle in 1960 of simple defined media that
chick embryo fibroblasts (left) and by poliovirus on HeLa cells, a line would support animal cell growth (i.e., not merely survival)
of cells derived from a human cervical carcinoma (right). Photo by
in culture. Together, these developments made possible the
R. Dulbecco; reprinted from S. E. Luria, J. J. Darnell, D. Baltimore, and
A. Campbell, General Virology, 3rd ed. (John Wiley & Sons, New York, quantitative study of animal viruses under carefully con-
NY, 1978), with permission. trolled conditions. The development of cell culture also made
possible the large-scale industrial production of vaccines,
including those for poliovirus, rubella, and measles virus.
Thus ends one chapter in our selective history. However,
procedure for animal viruses begins by removing the note that the elucidation of virus replication, as recounted
medium from the cell monolayer and then adding a small thus far, has not progressed beyond establishing the param-
aliquot of the virus sample to the cells. The virus is then eters of one-step growth. The story continues in the next
allowed to attach to the cells for as long as 1h, and the chapter, where we consider the biochemical and molecular
monolayer is then covered with a nutrient medium- events that occur during the eclipse period and the steps that
containing agar overlay. The purpose of the agar is to pre- go into generating progeny virus particles. We end the next
vent the diffusion of progeny virus beyond immediately chapter with a brief consideration of the variety of viruses
adjacent cells. Plaques that develop can generally be visu- and how we might sort them into families based upon the
alized by the addition of a vital dye that is concentrated organization and expression of their genomes.

Box 1.11
Phage and the Origins of Molecular Biology (J. Cairns, to study medically important viruses. The patient agreed,
G. S. Stent, and J. D. Watson [ed.], 1966) contains a collec- and since virology at Caltech was headed by Delbrck, the
tion of essays written in tribute to Max Delbrck by some of former physicist found himself with an endowment to study
the pioneers of molecular biology and virology who came human viruses, with no knowledge for how to use it. Well,
under his influence. Renato Dulbecco was one of those Delbrck summoned Dulbecco, an Italian physician who
scientists, and his essay relates how he came to work on joined the Caltech phage group to learn virology, to his
animal viruses. It happened as follows. office, and he proposed that Dulbecco give animal viruses
a try. Dulbecco was delighted by the idea, and the first of
In the late 1940s, a wealthy Californian became ill with his significant accomplishments (for which he later was
shingles (a late complication of chicken pox, actually caused awarded a Nobel Prize) was to develop a plaque assay proce-
by a herpesvirus, varicella-zoster virus; see Chapter 19). dure for animal viruses. Thus, quantitative animal virology
The mans physician explained that nothing could be done came to be.
for his shingles, and moreover, that virtually nothing was
known about the viruses that infect humans. Auspiciously, Dulbecco remarks on the importance of the development of
the physician knew of the studies being done on bacterio- the plaque assay for animal viruses, noting that subsequent
phages at Caltech, and he also was aware that Caltech was biochemical and molecular studies would have been much
the great center for such work. He explained to his patient less meaningful without reference to the multiplication
that bacteriophages were only of theoretical interest regard- phase revealed by the plaque assay. Interestingly, although
ing human disease, but he also suggested that the patient this was apparent to the phage workers, it was not equally
might help to develop a center at Caltech that might begin obvious to most animal virologists of the day.
History of the Nature of Viruses 19

Finally, before leaving the present chapter, we note that understanding of the regulation of cell growth and the
the study of animal viruses will come to incorporate, as molecular events that can lead to unregulated cell growth
well as lead to the development of, cutting-edge molecu- and cancer.
lar approaches. Moreover, these molecular approaches,
as well as studies of animal viruses per se, will come to Suggested Readings
have a dramatic impact not only on our understanding Cairns, J., G. S. Stent, and J. D. Watson (ed.). 2007. Phage and the
of viruses but also on our understanding of eukaryotic Origins of Molecular Biology: the Centennial Edition. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY. (This is the
cell and molecular biology as well. Research areas that latest edition of the original 1966 publication.)
will be impacted include eukaryotic gene organization, Girard, M. 1988. The Pasteur Institutes contributions to the
expression, and replication; intracellular signaling; en- fields of virology. Annu. Rev. Microbiol. 42:745763.
docytosis; and intracellular transport. Virus-centered Luria, S. E., J. E. Darnell, D. Baltimore, and A. Campbell. 1978.
studies also will come to have a major impact on our General Virology, 3rd ed. John Wiley & Sons, New York, NY.

Biosynthesis of Viruses:
Phage Protein Synthesis
RNA Metabolism in Infected Cells
Assembly of Progeny Phages
Packaging DNA within the Phage Particle
Modified Bases
Regulated Gene Expression
Phage Release: Lysozyme and the rII Region
Virus Classification
Animal Virus Structure
One purpose of this chapter is to examine the biosynthetic events that oc-
Entry of Animal Viruses cur during the eclipse period of a viruss replicative cycle. For historical
THE FAMILIES OF ANIMAL VIRUSES: reasons, and for the sake of continuity, we continue using the T-even
PRINCIPLES OF CLASSIFICATION phages as our model system (Box2.1). In doing so, we come upon addi-
Viral Genetic Systems: The Baltimore tional instances of virus-based research that led to breakthroughs that
Classification Scheme were of fundamental importance to biology in general. Since a major
purpose here is to learn about features of viral replication that are rather
general, we postpone until later in the chapter examining some aspects of
T-even phages that are unique to these viruses.
Also, we consider in some detail the means by which the temperate
bacteriophage (lambda) is able to follow either of two courses upon in-
jecting its DNA into a host cell; one course leads to virus replication and
cell lysis, and the other leads to lysogeny, which is the kind of latent infec-
tion manifested by temperate bacteriophages. Latent infection also occurs
among the animal viruses. Indeed, it is a defining characteristic of infec-
tions involving the herpesviruses. However, the establishment and main-
tenance of viral latency are best understood in the case of phage.
The chapter concludes with an introduction to the animal viruses and
a rational basis for the classification of animal viruses into families. But
first, we get on with our consideration of the T-even phages.
The T-even phages were the best-studied model system of the day (the
1940s). Yet their choice as the model virus was largely a matter of chance.
Although they were, at the time, thought to be the simplest possible or-
ganisms, the T-even phages are, in fact, among the largest, most complex
viruses that are known, containing about 160 genes. In contrast, there are
viruses that contain as few as five genes. Ideally, one would rather have
simple model systems. Yet, coincident with their complexity, the T-even
phages were found to have an unexpected feature that would lead to im-
portant new insights. That feature was the presence in their DNA of the
unusual base hydroxymethylcytosine (HMC) in place of cytosine (see the
chemical structure of HMC, below).
Biosynthesis and Classification of Viruses 21

Box 2.1
Where did the T-even phages come from, and what is the T-even phages, takes control of transcription in the infected
significance of their names? We may never know the exact cell, thereby converting the cell into a factory dedicated to
origin of these viruses, but the following is known. In the producing progeny phages. But whereas the T-even phages
late 1930s, T. L. Rakieten presented M. Demerec and U. Fano modify the host RNA polymerase to recognize phage-specific
with either a mixture of raw sewerage or perhaps a lysate promoter sequences (see the text), T7 encodes its own T7
from E. coli infected with raw sewerage. These two research- promoter-specific RNA polymerase, while also inactivat-
ers then isolated T3, T4, T5, and T6, all of which are active ing the host RNA polymerase. Because of the specific-
in E. coli. The precise origin of T2 phage, the most widely ity of the T7 RNA polymerase for T7 promoters, the T7
studied of the T phages, is less clear, although there is evi- polymerase and promoters have been incorporated into
dence that a researcher named J. Bronfenbrenner may have a number of cloning/expression vectors used in recom-
been working with T2 phage as early as 1932, and indeed binant DNA technology. T7 also encodes its own DNA
Bronfenbrenner may have initially isolated the virus. If that polymerase, a modified form of which is marketed as the
is the case, then it is likely that the isolation was made from popular sequencing enzyme Sequenase. Phage T1, which is
fecal material rather than from sewerage, as that appears to not related to any of the other T phages, remains the least
have been Bronfenbrenners practice. At any rate, Delbruck studied of the group.
named these E. coli phages Type 1 (T1), Type 2 (T2), Type 3
(T3), etc., for easy reference. Thus, the T-even phages are T2, The T-even phages, together with bacteriophage (see the
T4, and T6, as opposed to T1, T3, T5, and T7. text), have been the most intensely studied of the bacterio-
phages. While the T-even phages played a key role in the de-
Members of the T series of phages differ in size and shape. velopment of virology and molecular biology (as recounted
But, fortuitously (regarding their grouping as the T-even in Chapter 1 and the present chapter), bacteriophage
phages), phages T2, T4, and T6 are structurally similar, con- served as the prototype of the lysogenic (or temperate)
taining an extra row of subunits, which distorts the simple phages. Intense research on the T-even phages was largely
icosahedral structure. In addition, they contain similar tail due to the influence of Delbruck, who in 1944 urged the
structures and fibers, and they are unique in that they con- phage group to focus its research on the T series of phages.
tain the modified base 5-HMC instead of cytosine (see the Delbrucks rationale was to getall the phage researchers to
text). Moreover, they are related serologically and all have work on the same small group of phages, so that they might
large genomes. Thus, it is biologically meaningful to refer to more easily compare their experimental findings, in that
this group of viruses as the T-even phages. way facilitating progress. This recommendation of Delbruck
was considered correct at the time but has since been criti-
What then of the T-odd phages? These phages fall into three cized because it led to the early neglect of some other key
serological groups, in which phages T3 and T7 are related, issues, such as lysogeny. In that regard, none of the T series
but phages T1 and T5 are each distinct. Phage T7, like the of bacteriophages are temperate (see the text).

NH2 the 1950s demonstrated the presence of deoxycytidylate

hydroxymethylase, and other new enzymes, in T-even
N OH phage-infected cells.
O N These new enzymes in fact are encoded by phage genes,
H as was demonstrated at the time by the isolation of phage
mutants that are defective for their expression. While the
In addition, the HMC residues on the T-even phage finding that phages encode enzymes might not appear to
DNA are glucosylated in a specific pattern. be particularly remarkable today, at the time of this dis-
Since this form of modified DNA does not exist in un- covery it was quite striking. To appreciate why this was so,
infected cells, it was reasonable to infer that the phages we need to appreciate how little was known at that time
somehow are responsible for generating it. Taking this ar- about viruses, and of molecular biology in general. In this
gument a step further, one might hypothesize that these regard, recall where we are in our historical account of
phages cause the formation of new enzymes that previ- virology and molecular biology. Indeed, in our historical
ously did not exist in the cell. Indeed, Seymour Cohen in account of virology and molecular biology, we are only

slightly past the point at which DNA was finally acknowl- Head internal components
edged to be the genetic material and Watson and Crick IP, I, II, III, alt,
24 peptides II, VII,
had solved its structure. There was little else. soc DNA
The modified T-even phage DNA helped to advance hoc
the field in yet another important way. Its hydroxymethy- 23
lated glucosylated DNA could be distinguished from host
cell DNA, thereby making it possible to follow phage DNA
replication during infection of the host cell. 13, 14, N2, 4, 6 20
3, 15 wac
Internal tube, 19
36 18
Next, we briefly revisit the structure of T-even phage par- 34
ticles to describe in somewhat more detail how phage
DNA enters into the cell. We recall from Chapter 1 that
electron micrographs of the T-even phages showed that
they are tadpole shaped, with distinct heads and tails. In 9
addition, they appeared to attach to the outside of their 12
host cells by their slender tails. Base-plate wedge Base-plate hub
A more careful examination of T-even phage structure 6, 7, 8, 25, 53 5, 26, 27, 28,
frd 29, 48, 54, td,
shows the presence of tail fibers, which actually mediate
attachment of the phage to the host cell (Fig. 1.4, 2.1, and
2.2). Experimental proof that the tail fibers mediate Figure2.2 Diagrammatic representation of the structure of bacterio-
phage T4, with the locations of the known viral structural proteins.
phage attachment was obtained by isolating phage mutants Compare this to Fig. 2.1. Contributed by F. Eiserling; modified from
unable to generate tail fibers and then demonstrating that J. A. Levy, H. Fraenkel-Conrat, and R. A. Owens, Virology, 3rd ed.
these mutants are not able to attach. (A puzzle: if the phage (Prentice Hall, Englewood Cliffs, NJ, 1994).
mutant is defective for attachment and infection, how

can it be isolated and propagated for further study? The

Figure2.1 Electron micrograph of a T2 phage. Note the polyhedral answer is that the mutant is conditionally defective. Most
capsid, the neck, the contractile sheath, the base plate, and the tail
fibers. Compare this to Fig. 2.2. Photo by R. C. Williams; reprinted
such conditionally defective mutants contain temperature-
from J. A. Levy, H. Fraenkel-Conrat, and R. A. Owens, Virology, sensitive lesions. Although these mutants express their phe-
3rd ed. (Prentice Hall, Englewood Cliffs, NJ, 1994). notype at their restrictive temperature, they can be propa-
gated at their permissive temperature.) In addition,
attachment was impaired by antibodies reactive against
the tips of the tail fibers, but not by antibodies reactive
against other sites on the virus particle.
The phage tail fibers attach specifically to lipopolysac-
charide receptor molecules on the outer envelope of the
host cell. This binding event triggers a conformational
change or bending of the tail fibers that brings the base
plate at the end of the tail into contact with the cell surface
(Fig. 2.3). Ensuing conformational changes then occur in
the base plate and in the tail. Note that the tail is comprised
of an outer sheath that surrounds a central tube (Fig. 2.2).
The outer sheath of the tail contracts, and the hole in the
center of the base plate enlarges. The contraction of the
sheath drives the tube through the enlarged hole in the base
plate, and ultimately through the cell wall. The process is
likely facilitated by a viral enzyme (a lysozyme), built into
the phage tail, which degrades the cell wall from the out-
side. The phage DNA is now able to pass from the phage
head through the tube and into the cell (Box2.2).
Biosynthesis and Classification of Viruses 23

Figure2.3 Diagrammatic representation of bacteriophage T4 attaching to a bacterium and injecting its DNA into
the cell. (A and B) The phage tail fibers specifically bind to receptor molecules on the outer envelope of the bacterium.
(C) This triggers a conformational change or bending of the tail fibers, which brings the base plate at the end of the
phage tail into contact with the bacterial surface. (D and E) Here the tail is shown comprised of an outer sheath that
surrounds a central tube. Conformational changes occur in the base plate and in the tail. The outer sheath of the tail
contracts, and the hole in the center of the base plate enlarges. The contraction of the sheath drives the tube through
the enlarged hole in the base plate and ultimately through the cell wall. A phage lysozyme digests away the cell below
the phage, enabling the phage DNA to pass from the phage head through the tube and into the cell.

Box 2.2
What force might drive the phage DNA from the phage head and, in many instances, through a long tail as well. What
into the cell? Ever since Hershey and Chase showed that T2 then drives the complete release of the viral genome? The
phage inserts its DNA into the cell, it has been assumed that answer is not known in most cases. But studies of phage
the pressure exerted by the DNA within the capsid (now T7show that the majority of its genome is actually drawn
known to be as much as 60 atmospheres) provides the driv- from the capsid by transcription. The first portion of the
ing force for injection and release of the DNA. Comparable T7 genome, about 850 bp, enters the cell by a transcription-
pressures should exist within the capsids of other phages as independent process, perhaps driven by pressure exerted
well. However, consider that during bacteriophage assembly, by DNA within the capsid. Then, about 7 kilobases (kb)
phage genomes generally are inserted into preassembled of DNA is drawn out by transcription catalyzed by the
empty capsids (see the text). Next, consider the following E. coli RNA polymerase, and the following 32 kb is drawn
experimental finding from studies involving the encapsida- out by transcription catalyzed by the T7 RNA polymerase.
tion of phage j29. Half of the j29 genome can be encapsi- The initial transcription-independent translocation phase
dated before substantial forces that resist further packaging exposes the gene encoding T7 RNA polymerase, enabling
become evident. Consequently, during the reverse process of the phage enzyme to be produced. Phage N4 also is known
DNA release, internal forces within the capsid should have to harness transcription to facilitate its entry. However, we
largely dissipated after only about one-half of the genome can be rather certain that there are a variety of strategies
has been released. Also, bear in mind that release of the phage used by different bacteriophages to release their genomes
genome involves passage through a head-tail connector into the cell.

The mechanism by which T-even phages deliver their encloses an internal tube), base plate, and tail fibers. The
DNA into host cells highlights the fact that virus particles head, collar, and base plate each are assembled from multi-
indeed may be complex molecular machines, rather than ple copies of several individual proteins. Each of the tail
mere inert chemical configurations (Box2.2). Examining fibers is also composed of several different proteins. The re-
the structure of T-even phage particles in somewhat more markable outer sheath is assembled from 144 molecules of
detail and noting the number of their distinct protein a protein that is arranged in 24 rings. When the tail fibers
components underscore this point (Fig. 2.2). For example, attach to the cell, a change in angle between their compo-
consider the particles distinct head, collar, sheath (which nents stimulates subsequent molecular reorganization,

including contraction of the sheath, a process that involves considerable amount of this phage-specific protein syn-
a complex rearrangement of the sheath proteins from the thesis occurs before phage DNA replication commences.
initial 24 rings into 12 rings. These points will be considered in more detail shortly.
What might be the functions of the early phage pro-
teins that are produced before phage DNA replication?
SEQUENCE OF PHAGE BIOSYNTHETIC Might any of these proteins be structural components of
EVENTS the phage particles? This specific question was answered
Phage Protein Synthesis by first preparing antiserum against the proteins that con-
Transport of the phage DNA into the cell marks the start stitute phage particles, which was done by simply inject-
of the eclipse period. Thus, we are now ready to consider ing rabbits with disrupted phage particles. The resulting
the succession of biosynthetic events that result in the antiphage antiserum was not reactive against the early
production of progeny phage (Fig. 2.4). We begin by not- phage proteins, thus showing that those phage-encoded
ing the marked increase in protein synthesis that begins proteins that are made prior to phage DNA synthesis are
almost immediately upon entry of the phage DNA. Fortu- not components of phage particles. In contrast, the late
itously (for the experimenter), the phage shuts down host phage proteins, made after the phage DNA replicates, are
protein synthesis soon after DNA entry. Thus, the protein reactive with the antiphage antiserum, indicating that
that accumulates in the infected cells is virus specific. Not they include structural components of phage particles.
all viruses shut down host protein synthesis. Nevertheless, These experimental findings demonstrate that there are
the redirection of protein synthesis from cellular products two functionally distinct classes of T-even phage proteins,
to phage products in this instance points to a fundamental whose expression is separated in time, and which are re-
feature of viral parasitism: the subversion by the virus ferred to as early and late proteins. Stated somewhat dif-
of the hosts biosynthetic machinery. Also, notice that a ferently, there is a program of regulated phage gene ex-
pression. These points will be considered in more detail
shortly. First, we need to answer the question we began
Figure2.4 Schematic representation of the pattern of biosynthesis in E. with: what is the function of the early proteins?
coli infected with bacteriophage T4. The broken lines in the bottom part The HMC in the phage DNA provides a clue as to the
of the figure depict the pattern of phage early protein synthesis in bacteria
infected with conditional phage mutants that are unable to initiate phage
function of the early phage proteins. As argued above,
DNA synthesis. Virion refers to a virus particle. Adapted from S. E. since HMC does not exist in uninfected host cells, its
Luria, J. J. Darnell, D. Baltimore, and A. Campbell, General Virology, presence in infected cells implies the existence of phage-
3rd ed. (John Wiley & Sons, New York, NY, 1978), with permission. encoded enzymes that might function in phage nucleic
acid metabolism. Seymour Cohen and Arthur Kornberg
demonstrated that indeed this is the case, showing that a
Total protein group of different phage-encoded enzymes need to be
3 Phage DNA
Virion protein
made before phage DNA synthesis can begin. These
Arbitrary units

Infective virions 1950s experiments were based on genetic analysis. (Later

2 T4 experiments, done by others, involved invitro translation
systems prepared from the Escherichia coli host cells, in
which protein synthesis was directed by added phage-
1 Total RNA specific messenger RNA [mRNA] that had been tran-
scribed invitro. However, from our historical perspective,
Bacterial RNA mRNA still remains to be discovered, so these experiments
cannot yet be a part of our story.)
3 The early phage proteins include (i) enzymes that cata-
Phage lysozyme
lyze the synthesis and glucosylation of HMC, (ii) enzymes
Arbitrary units

2 Early phage that destroy precursors of cytosine deoxynucleotides,

enzymes (iii) enzymes that synthesize thymine by a new pathway,
(iv) enzymes that degrade host cell DNA to make its nu-
1 Bacterial enzymes
cleotides available for phage DNA synthesis, and (v) a
phage-specific DNA polymerase. The roles of some of these
enzymes in phage DNA metabolism will be considered in
0 10 20 30 40 more detail shortly. We briefly note here that they enable
Minutes the phage to block any further cellular DNA synthesis and,
Biosynthesis and Classification of Viruses 25

moreover, they enable the phage to degrade cellular DNA. this fact is considered later in the chapter. For now, the
The nucleotides that are released from the degraded cel- appearance of intracellular progeny phage marks the end
lular DNA are then available to the phage to use as build- of the eclipse period and the onset of the intracellular rise
ing blocks for phage DNA replication (Box 2.3). Yet the period.
two essential points for the moment are, first, that the
phage DNA directs protein synthesis before the phage RNA Metabolism in Infected Cells
DNA replicates and, second, that these early phage-specific We now turn our attention to the pattern of RNA metabo-
proteins are not structural components of phage particles lism in T-even-phage-infected cells. To appreciate what will
(virions) but rather are enzymes. next unfold, note that scientists of the late 1950s still knew
Once phage DNA has been replicated, phage structural virtually nothing about the role of RNA. If we can imagine
component proteins then begin to appear (Fig. 2.4). An- for the moment that we too know virtually nothing about
other important feature of the phage replication cycle is the role of RNA, then we can better appreciate how the
that these phage structural components appear before the study of phage-infected cells led to breakthroughs of pro-
emergence of progeny phage particles. The significance of found significance to virology and to biology in general.

Box 2.3
What follows is an example of the T-even phages serving ultracentrifuge, the heavy metal ions form a concentration
to test a hypothesis of paramount importance. Watson gradient and consequently a density gradient. At equilib-
and Crick in 1953, in their very short paper reporting the rium, experimental samples applied to the heavy metal
structure of DNA, succinctly stated that the DNA structure salt solution would be found in the density gradient at the
in fact suggested a mechanism for its replication. Interest- positions of their buoyant densities. In this way, heavy DNA
ingly, they did not elaborate on what that mechanism might might be separated from light DNA in the density gradient.
be. But what they certainly had in mind was a mechanism
in which the two strands of the parental double helix might In one version of their experiment, Meselson and Stahl
separate, with each strand then serving as a template for proceeded as follows. They grew a batch of phage in the
the formation of a new complementary strand, assembled presence of radioactive 32P, so that the input parental phage
from a pool of purine and pyrimidine bases. This process DNA strands would contain a radioactive label. They then
is referred to as semiconservative replication, since each used these phage, which contained 32P-labeled DNA, to in-
daughter DNA molecule would contain one old and one fect cells that were growing in the presence of 15N and 13C as
new polynucleotide chain. the sole nitrogen and carbon sources, respectively. The DNA
from the progeny phage issuing from this infection was then
Well, regardless of how attractive this hypothesis might have analyzed by equilibrium density-gradient sedimentation.
been, it still needed to be verified experimentally before it
could be accepted. In a classic and elegant experiment that So, if Watson and Crick are correct, and DNA replication
was inspired in part by the finding that T-even phages shut is semiconservative, where would you expect to find the 32P
down cellular macromolecular synthesis, Matthew Meselson label in the density gradient? Would it be with the heavy,
and Franklin Stahl in 1958 used T-even phages to dem- dense DNA or with the light DNA? Answer: It was found in
onstrate that DNA replication in fact is semiconservative. a band of hybrid density, between the heavy and the light
Their experiment made use of the fact that the rare isotopes bands. Can you explain this finding? Does it prove that
N and 13C are heavier than the common isotopes 14N and DNA replication is semiconservative?
C. Thus, DNA that incorporated the heavier isotopes
would be denser than DNA containing the lighter isotopes. Note that Meselson invented equilibrium density-gradient
Moreover, heavy DNA could be separated from light centrifugation using cesium chloride, which made the
DNA by equilibrium density-gradient sedimentation in an experiment possible. Meselson was a graduate student at the
ultracentrifuge. time, and Stahl was a postdoctoral student, each at Caltech.
Interestingly, Meselson first met Stahl at the Marine Biologi-
In the centrifugation procedure, samples are suspended cal Laboratory at Woods Hole in the summer of 1954, when
in a salt solution of a heavy metal, such as cesium. Under Meselson was a teaching assistant in James Watsons physiol-
the influence of the centrifugal force generated in the ogy course and Stahl was a student in the course.

We begin by observing that there is a slight increase in polynucleotide chains that are complementary to each
total RNA after the start of infection and that the level of other in their base sequences. The hybridized RNA could
RNA then appears to remain constant over time (Fig. 2.4). be detected by several means. For example, it is resistant
The apparently constant levels of RNA seen in infected to RNase. Also, filters that allow single-stranded RNA to
cells would seem to suggest that there is not much RNA pass retain the hybridized RNA. The same principle of
metabolism going on. However, this conclusion would be hybridization between complementary RNA and DNA
quite incorrect. A fraction of that RNA is actually contin- strands underlies more contemporary experimental ap-
ually being synthesized and degraded. This fact was dem- proaches, such as gene arrays.)
onstrated by a pulse-chase experiment, done as follows. Our next question, and a singularly important one at
P was added to the infected cells for a brief period of that, concerns the function of the unstable phage-encoded
time (the pulse), to label the RNA. The cells then were RNA. Bearing in mind the level of knowledge at the time,
washed and incubated in an excess of unlabeled precursor now the very early 1960s, it was reasonable to suggest that
(the chase). At various times the RNA was extracted, and this RNA has some function in protein synthesis, since a
the amount of radioactivity it contained was determined. general correlation between RNA levels and protein syn-
This experiment showed that the amount of label that was thesis had been apparent for several years. Moreover, it
incorporated into the RNA during the pulse continually was clear that DNA could not directly serve as the tem-
decreased during the chase, with a half-life on the order of plate for protein synthesis. This point was obvious from
minutes, thus demonstrating that this RNA rapidly turns the example of eukaryotic cells, in which the DNA is
over. That is, this RNA fraction is unstable and is continu- contained within the membrane-bound nucleus, whereas
ally replenished. protein synthesis occurs in the cytoplasm. Moreover, cy-
A major question was whether this rapidly turning toplasmic ribosomes were known to be the sites of pro-
over RNA is cellular or viral in origin. An early experi- tein synthesis. (This was shown by a pulse-chase experi-
mental approach to answer this question was to ascertain ment that was done as follows. First, radioactively labeled
whether the base composition of that RNA might reflect amino acids were injected into rats. Then, at various
the base composition of either the phage DNA or the cel- times rat liver tissue was removed and homogenized,
lular DNA. This was done in 1956 by Volkin and Astra- and cell fractions were prepared and analyzed for their
chan, who demonstrated that the unstable RNA in T2 content of the radioactive label. Within 2min, radioac-
phage-infected cells has the same base ratio (A+U/G+C) tively labeled polypeptides were observed in the ribo-
as the phage DNA (with T in place of U), rather than that some fraction. Radioactively labeled proteins then were
of the cell DNA. This experimental finding strongly im- found loose in the cell sap [i.e., the liquid phase of the
plied that this rapidly turning-over RNA is phage spe- cytoplasm]. Thus, these experimental results showed
cific, rather than cell specific, and, importantly, that it that amino acids are assembled into proteins on ribo-
might be encoded by the phage DNA. Proof of this fact somes, and the completed proteins then are released into
came from Hall and Spiegelman in 1961, when they found the cytoplasm.)
that this RNA could be hybridized with phage DNA, but Since ribosomes were known to contain RNA, it was
not with cell DNA. (The principle behind the approach reasonable to suggest that the intermediary in protein
developed by Spiegelman is explained here for the benefit synthesis might be the RNA component of the ribosomes.
of those to whom it might be new. Heating double- This hypothesis, although attractive, would prove to be
stranded DNA molecules to temperatures above their wrong in a subtle but important way. It implied that each
melting point [just below 100C] causes the hydrogen gene is transcribed to a unique RNA, which in turn is in-
bonds between the purines and pyrimidines of the two corporated into a specialized ribosome. That specialized
complementary strands to break, thereby enabling the ribosome could then direct the synthesis of only one par-
strands to separate. If this denatured DNA then is slowly ticular protein. If that idea were correct, then we might
cooled [annealed] to room temperature, the complemen- expect the rapidly turning over phage-encoded RNA to be
tary strands reestablish their interchain hydrogen bonds, associated only with the new ribosomes that are made af-
and DNA double helices re-form. If single-stranded RNA ter the start of infection. So then, what is the relationship
molecules are added to the denatured DNA before the between the phage-encoded RNA and the ribosomes? Spe-
slow-cooling step, and if that single-stranded RNA is cifically, is that RNA associated only with new ribosomes?
complementary to some of the DNA strands, then RNA- In a 1961 experiment notable for its logic and elegance,
DNA hybrid helices will form, in addition to the DNA- in addition to its exceptional importance, Sydney Brenner,
DNA helices, when the cooling occurs. Importantly, the Francois Jacob, and Matthew Meselson established the re-
hybrid molecules will form only between RNA and DNA lationship between the rapidly turning over, phage-encoded
Biosynthesis and Classification of Viruses 27

RNA and ribosomes. If the one geneone ribosomeone corollary principle is that once a virus has managed to
protein hypothesis was correct, and if the unstable phage- generate mRNAs, the cellular translation machinery is
specific RNA played an intermediary role in protein syn- available to the virus to translate that viral mRNA into
thesis, then the phage-specific RNA should be found as- protein. As we will see shortly, this concept provides the
sociated only with ribosomes that are made after infection basis for a rational scheme for classifying all of the virus
is under way. Thus, Brenner, Jacob, and Meselson needed families into a few groups, based on the relatively few
to distinguish the ribosomes made after infection had be- strategies by which the many different virus families tran-
gun from the ribosomes that were present before infec- scribe their genomes.
tion. They did this by growing cells for several generations
in medium containing the heavy isotopes 15N and 13C as Assembly of Progeny Phages
the sole respective nitrogen and carbon sources. In this As noted above, the structural protein components of the
way, essentially all ribosomes present in the cells before phage particles are not made until after the onset of phage
infection would be heavy. Next, the cells were washed DNA synthesis (Fig. 2.4). This pattern of regulated late
and placed in medium containing only the common iso- gene expression is a characteristic feature of the replica-
topes 14N and 12C. Moreover, the cells were immediately tion of DNA viruses in general. Soon, we will consider the
infected with the phage upon transfer to the fresh me- advantage that this pattern of regulated gene expression
dium. In that way, any new ribosomes made after infection might confer upon these viruses. We also will consider
was under way would be light. the mechanisms that underlie it, both in the case of the
Importantly, recall that the phage-encoded RNA could T-even phages and, in later chapters, in the cases of other
be specifically labeled by adding 32P to the infected cul- viruses as well. But first we need to consider two other
tures. (You may have been wondering why the 32P is incor- general characteristics of virus replication that are revealed
porated only into phage-encoded RNA. The reason will by the analysis of the T-even phage replication cycle. The
become clear shortly.) Thus, 32P was added to the cultures first of these is the accumulation of phage structural
after infection to label any de novo phage-encoded RNA. components in the cells, before the emergence of progeny
Now, Brenner and coworkers needed only to determine phage particles, the significance of which is demonstrated
whether the 32P-labeled phage-encoded RNA was associ- by the phenomenon referred to as phenotypic mixing, as
ated only with new light ribosomes. Heavy and light ribo- in the following example.
somes could be separated by equilibrium density gradient Before we can do our actual experiment that demon-
centrifugation. (In 1958, Meselson and Frank Stahl used strates phenotypic mixing, we first need to isolate particu-
T-even phages and equilibrium density gradient centrifu- lar sorts of cell and phage mutants. First, the cell mutants.
gation to demonstrate that DNA replication in fact is Recall that T2 phage initiates infection by specifically
semiconservative, as recounted in Box 2.3.) When all of binding via its tail fibers to receptor molecules on host
this was done, the 32P-labeled phage-encoded RNA was bacterial cells. By exposing a culture of susceptible E. coli
found in association with the old heavy ribosomes that B cells to the phage, we can select for bacterial mutants
were made before infection. Indeed, there were no new that survive exposure to the phage. These T2-resistant
light ribosomes in the infected cells, since the phage ef- bacteria, which are referred to as E. coli B/2, have muta-
fectively shuts down cellular macromolecular synthesis. tions that alter the receptor for the phage, so that the
The result of this experiment is of crucial importance phage can no longer attach to them.
to biology in general, since it established the existence of Next, we select for phage mutants that have the ability
mRNA, which carries the information encoded in the to attach to, and infect, the T2-resistant E. coli B/2 cells.
gene to unspecialized ribosomes, which already contain This is done simply by exposing a culture of E. coli B/2
their own general ribosomal RNA (rRNA). The unstable cells to a high-titer inoculum of the T2 phages. The phage
mRNAs enter into transient associations with the ribo- mutants that are selected in this way are referred to as
somes, where they are translated into the proteins that T2h, where the h refers to the altered host range of the
they encode. Thus, rather than one geneone ribosome mutants. As you might predict, the mutation in the T2h
one protein, we now have one geneone mRNAone mutants affects their tail fibers, enabling them to bind to
protein. These experiments soon led to the successful the altered receptors on the E. coli B/2 cells. Some of these
search for an analogous cell-specific mRNA fraction. T2h mutants also retain the ability to infect the original
The experimental results of Brenner, Jacob, and Mesel- E. coli B cells.
son are also of fundamental importance to virology, since We can now proceed with our actual experiment that
they establish that viruses subvert the host cell protein demonstrates phenotypic mixing. First, we expose E. coli
synthesis machinery to translate their own proteins. A B cells to an inoculum that contains a high titer of both T2

and T2h phages. In this way, every cell is infected with actual sequence of steps by which structures as complex as
both wild-type (normal) and mutant phages. Second, we these phages are assembled. The facts that phage particles
harvest the lysate from this mixed infection and use it as are assembled from a pool of precursor components and
the inoculum to infect a fresh culture of E. coli B/2 cells. that those individual components are taken at random
Third, we analyze the virus in the lysate from the infected from the pool do not mean that assembly is a random
E. coli B/2 cells. We might have expected that this lysate process. Instead, assembly occurs via a specific sequence
would contain only T2h mutants, since wild-type T2 of steps. R. S. Edgar and W. B. Wood analyzed the T4 phage
phage would not have been able to infect the E. coli B/2 assembly pathway in 1965. They made extensive use of a
cells. (We can easily distinguish T2h mutants from the variety of phage mutants, each of which is unable to pro-
wild-type T2 phage by differences in the plaques [see duce infective phage particles. These mutants were used in
Chapter 1] that they each produce on a mixed culture of two different ways. In their first experimental approach,
E. coli B and E. coli B/2 cells. T2h mutants lyse both E. coli Edgar and Wood infected E. coli cells with the individual
B and E. coli B/2 cells to produce a clear plaque. In con- mutants and then determined by electron microscopy the
trast, wild-type T2 phage lyses only the E. coli B cells. Con- assembly intermediates that accumulated in each instance.
sequently, the wild-type T2 phage produces a turgid Their second and especially innovative approach put to
plaque on the mixed culture of E. coli B and E. coli B/2 use their observation that certain combinations of ex-
cells.) We find that, in fact, the lysate from the infection of tracts from mutant-infected cells actually complemented
the E. coli B/2 cells contains wild-type T2 phage as well as each other invitro, thereby generating infective particles.
the T2h mutants. So then, what is the explanation? Before From their extensive electron microscopy analysis of a
reading on, try to solve this for yourself. Congratulations large number of individual viral mutants together with
if you deduced the answer, which is as follows. their in vitro complementation experiments, Edgar and
Virus particles are assembled from the pool of compo- Wood deduced a rather complete picture of the T4 assem-
nent protein molecules that accumulate in the infected bly pathway (Fig. 2.5). In particular, note the three inde-
cells. When the mutant and wild-type phages are grown pendent branches leading to the formation of heads, tails,
together in the same cell, as in the first step of our experi- tail fibers, and then complete phage particles.
ment, they each contribute to the pool of phage compo- As discussed shortly, bacteriophages and animal vi-
nent proteins. When progeny phage particles are assem- ruses have evolved different structural features to solve
bled from the pool of phage proteins, these proteins are the particular obstacles that they face in gaining entry into
withdrawn from the pool at random, without regard to their respective host cells. Despite these differences, im-
the genotypes of the phages that produced them. Thus, in portant general principles of virus assembly are revealed
the first step of the experiment, in which E. coli B cells by the T-even phages. Foremost of these is that virus par-
were infected with both T2 and T2h phages, progeny ticles are assembled from multiple copies of identical pro-
phages were assembled that contained either a T2 or a T2h tein subunits. Second, viruses use subassembly pathways
genome and a mixture of protein components encoded by rather than a single linear pathway. Significantly, these two
both the wild-type T2 and mutant T2h genomes. These principles provide for the limited number of viral genes to
phages are referred to as phenotypically mixed. With be efficiently utilized and for the possibility of rejecting
that in mind, and recalling that the tail fibers confer host mistakes before the entire particle might be ruined.
range specificity, many of the phages generated in the first
step of the experiment thus contain wild-type T2 genomes Packaging DNA within the Phage Particle
in phage particles with some T2h tail fibers (there are six The encapsidation of the viral genome within the protein
tail fibers inall). Some of these phages have enough T2h head, or capsid, is a crucial step in the assembly of DNA
tail fibers to enable them to infect E. coli B/2 cells in the viruses that still is poorly understood. One might have ex-
second step of the experiment. However, any phage par- pected that the simplest way to package the phage DNA
ticle that has a wild-type genome will still produce a tur- would be to assemble phage capsids around a collapsed
gid plaque on the mixed culture of E. coli B and E. coli B/2 form of the DNA, since that would avoid the energetically
cells. This example of phenotypic mixing demonstrates unfavorable process of inserting the DNA molecules, with
the following characteristic aspect of the viral life cycle. their negatively charged phosphate groups, into preassem-
Virus particles are assembled from a pool of precursor bled empty phage heads. Regardless, pulse-chase exper-
components. This is fundamentally different from the cel- iments demonstrated that empty phage proheads indeed
lular lifestyle, in which cells grow larger and then divide. are precursors in the assembly pathway. That is, preas-
The remaining general characteristic of viral replica- sembled empty proheads are filled with phage DNA. To
tion that is revealed by the T-even phages concerns the appreciate how remarkable this is, consider the following.
Biosynthesis and Classification of Viruses 29

Proximal half

Tail fiber 34, 57

37, 57 Distal half

38 36 35 63


20, 21, 22, 23, 16, 17, 49 2, 4, 50, 13, 14
24, 31, 40, 66 64, 65

Tail-head connector

9, 11, 12 48 54 19 18 3, 15
Tail pins Sheath
5, 6, 7, 8, 10, 25, 26,
27, 28, 29, 51, 53


Figure2.5 The assembly pathway of the T-even phages has three main branches that lead independently to the
generation of phage heads, tails, and tail fibers. These branches then converge to yield mature phage particles. The
numbers refer to the particular phage genes whose products are involved at each particular step. A mutation in any
of the indicated genes leads to the accumulation of the intermediate structure seen immediately before the step in
which the mutant gene is involved, as well as in the accumulation of the terminal structure of each of the other two
pathways (i.e., the structures just before the convergence of the three branches). The solid arrows indicate steps that
have been demonstrated invitro. Modified from W. B. Wood, in F. H. Ruddle (ed.), Genetic Mechanisms of Develop-
ment (Academic Press, New York, NY, 1973), as adapted by G. S. Stent and R. Calendar, Molecular Genetics, 2nd ed.
(W. H. Freeman and Company, San Francisco, CA, 1958), with permission.

The diameter of the phage head is 0.085 m. The length of structural proteins that make up the prohead. The confor-
the phage DNA is nearly 6,000 times greater at 50 m mational rearrangements of those proteins result in a
(Fig. 1.5). Moreover, phosphate groups on the outside of more stable conformation of the prohead, as it matures
the DNA double helices result in charge repulsion as the into the head. By combining the energetically favorable
DNA is threaded into the proheads. Notwithstanding conformational rearrangement of the protein lattice of
these facts, the DNA within the head is yet packaged to the head with the energy-requiring DNA packaging, the
liquid crystalline density. overall process can be more energetically favorable.
Despite the general lack of understanding of the T-even The prohead is able to undergo its conformational re-
phages encapsidation process, several aspects of the pro- arrangement because it is in a metastable conformation.
cess that are known are relevant to virus assembly in gen- This is an extremely important point, in this instance and
eral and also to the notion of viruses behaving as dynamic in general, since the release of a structure from its meta-
molecular machines. One such aspect is that proheads stable conformation and its subsequent assumption of its
contain, in addition to their permanent structural com- most stable conformation can provide the driving force
ponents, several different proteins that transiently asso- for subsequent events.
ciate with the proheads. These transient proteins exert From thermodynamic considerations, we might have
scaffolding and chaperone functions that facilitate the as- expected that the prohead should assemble directly
sembly of the proheads. They are eliminated from the into its most stable conformation. How, then, is the meta-
prohead as DNA encapsidation proceeds. Importantly, ex- stable state achieved? The general principle is to make use
clusion of the scaffolding proteins as the prohead fills with of specific proteolytic cleavages. In this instance, the
DNA is associated with rearrangements of the permanent multicomponent prohead structures are assembled from

protein subunits, some of which are later cleaved. The ini- unique to these particular viruses. The first of these fea-
tial structure indeed is in its most stable conformational tures is the presence in their DNA of the modified base
state. However, the structure becomes metastable upon HMC. An examination of how this strange base is gener-
cleavage of particular subunits, in this case including the ated and then incorporated into the phage DNA will en-
scaffolding proteins that also are extruded during encapsi- able us to appreciate how it serves the phages purpose.
dation of the DNA. First, a phage-encoded enzyme (deoxycytidine tri- and
Another detail of the T-even phage encapsidation pro- diphosphatase) dephosphorylates the intracellular deoxy-
cess, which also is seen in the cases of some animal viruses, cytidine triphosphate (dCTP) and deoxycytidine diphos-
is the presence of a portal complex located at one of the phate (dCDP) pools to deoxycytidine monophosphate
12 vertices of the phages icosahedral head structure (see (dCMP). At this point, DNA synthesis cannot occur for
below). That portal complex (also known as a connector want of dCTP as a precursor. Next, the phage-encoded
complex) is a dodecameric structure that contains a cen- enzyme deoxycytidylate hydroxymethylase adds a hy-
tral channel that is about 30 angstroms () in diameter. It droxymethyl (HM) group to dCMP to yield HMdCMP.
functions as the entranceway for the phage DNA into the Another phage-encoded enzyme, deoxynucleotide kinase,
prohead. Also, it is the channel through which the DNA phosphorylates HMdCMP, yielding HMdCDP, which it
exits from the head during infection. ATP hydrolysis in turn phosphorylates to yield HMdCTP. The virus-
provides energy for the not-yet-understood driving force encoded DNA polymerase then uses HMdCTP as a pre-
that moves the phage DNA through the channel into the cursor for the synthesis of progeny phage genomes, which
prohead. thereby come to contain HMC.
Just as the nature of the driving force for encapsidation The clever phage is here exploiting the important fact
is not clear, the driving force that causes the DNA to exit the that the phage DNA polymerase, but not the cellular DNA
particle at infection is not yet known either (Box2.2). One polymerase, can use HMdCTP as a precursor for DNA
possibility is that the pressure exerted by the DNA within synthesis. Since the phage has converted the intracellular
the head drives it out. In the case of phage j29 (phi 29), that pool of dCTP to a pool of HMdCTP, the phage has essen-
pressure has been estimated to be 6 megapascals (MPa). tially commandeered the precursor pool of nucleotide
The tensile strength needed to contain such a pressure is bases to its own purpose. In addition, the presence of the
equivalent to that exerted by contemporary aluminum al- strange base in the phage DNA protects it from a phage-
loys. Bear in mind that the molecular motor that packages encoded endonuclease that is specific for unmodified
the DNA into the head must work against that pressure. cytosine-containing host DNA. The nucleotides that are
The exceptional tensile strength of the phage head, and released from the degraded cell DNA are used by the
of virus capsids or shells in general, results from the strong phage as building blocks for its own DNA replication.
interactions between the viruss structural proteins. Fea- Moreover, glucosylation of the HMCs in the phage DNA
tures of virus structure that underlie these strong interac- by yet other phage-encoded enzymes (- and -glucosyl
tions are discussed shortly. Here we note that in addition transferase) protects the phage DNA from the defensive
to being able to sequester the phage DNA in the face of the cellular restriction endonucleases (Box 2.4). This is all
enormous pressure exerted by the DNA, the capsid also rather diabolical!
protects the DNA from extremes of temperature, pH, and Modified bases in biological nucleic acids are of
potentially lethal chemicals and radiation. Yet under ap- course found in cellular transfer RNA (tRNA). However,
propriate specific circumstances, which are recognized by large-scale replacement of a standard base with a modi-
the very same exceptionally stable capsid, it releases its fied one is found only in the DNA of certain bacterio-
DNA so that the next replication cycle can begin. The ex- phages. Other examples include the Bacillus subtilis
planation for this seeming paradox lies partly in the fact phages SPO1 and PBS1, which contain hydroxymethylu-
that even after the virus particles mature they still are in a racil instead of thymine and uracil instead of thymine,
metastable state. These facts further highlight a level of respectively. These strange bases in the DNA of the
functional elegance that virus particles have evolved, B. subtilis phages likely serve the same purpose as the
which we probably did not appreciate earlier. HMC in the T-even phages.

Regulated Gene Expression

UNIQUE FEATURES OF T-EVEN PHAGES The temporal regulation of T-even phage gene expression
Modified Bases is an important feature of the replication cycle of these
We now reconsider in somewhat more detail some special viruses, which was too briefly discussed above. We con-
features of the T-even phage replication cycle that are tinue by first considering the possible advantages to these
Biosynthesis and Classification of Viruses 31

Box 2.4 those structural component molecules is incorporated

into a progeny virus particle only once. Thus, it is logical
for the virus to begin by synthesizing the small amount of
Glucosylation, as the means by which T-even phages
the catalytic proteins necessary for it to amplify the num-
undermine host cell restriction, was demonstrated as fol-
lows. Phages with glucosylated DNA were used to infect
ber of viral genomes. Likewise, it is logical for the virus to
mutant E. coli B cells that were unable to make uridine wait to make its structural components until a much
diphosphate glucose (UDPG). Since UDPG is a necessary larger number of progeny DNA genomes have become
substrate for glucosylation of the phage DNA, its absence available to direct that process.
meant that progeny phage DNA could not be glucosylat- It also is advantageous to the virus to make very large
ed. However, phage DNA still could be replicated in the amounts of its structural components because these com-
UDPG-negative cells, probably because host restriction ponents must find, and interact with, each other in an in-
enzymes are membrane bound and act on the incoming tracellular milieu in which the concentration of cellular
phage DNA at entry. Also, at an early point in infection, proteins may be as high as that attained in protein crystals
the phage makes a protein that inhibits any cytoplasmic (i.e., 20 to 40 mg/ml). Since many viral structural sub-
restriction enzymes. The phage encodes such a protein units may inadvertently interact with irrelevant cellular
because new phage DNA is not immediately glucosylated. proteins, and since viral assembly is driven in the forward
When T2 phage with nonglucosylated DNA (referred to direction by the high concentration of viral structural
as T2*) was allowed to infect fresh E. coli B host cells, its subunits, it is beneficial to the virus to synthesize a great
DNA was rapidly degraded by restriction endonucleases excess of structural proteins; this can be done most effi-
present in those cells. However, T2* phage was able to ciently by first amplifying the genomic templates as de-
replicate in Shigella, which lacks the restriction enzymes. scribed above. (The fact that virus assembly takes place in
Interestingly, its DNA was glucosylated in the Shigella an intracellular environment in which the concentration
cells, so that it could then infect E. coli B. of irrelevant cellular proteins is strikingly high also may
account in part for the strategy of many viruses, animal
viruses as well as bacteriophages, to shut down cellular
protein synthesis.) A further and obvious advantage to the
virus of maximizing its capacity for directing protein
phages of their temporally regulated pattern of gene ex- synthesis is that it enables the virus to generate the maxi-
pression, and then we consider the mechanisms by which mal number of progeny virus particles.
it is accomplished. The production of large amounts of foreign viral pro-
It is logical that phage genes, which encode enzymes teins surely must be disadvantageous to the cell, particu-
required for phage DNA replication, are expressed prior larly since at least some of these proteins may promote
to DNA replication. However, what might be the benefit virus release by compromising the integrity of cellular
to the virus of postponing the expression of genes that membranes. Other virus proteins may compromise other
encode structural and other proteins that are required for cellular processes that support virus replication. Thus, by
generating phage particles, until after the phage DNA has synthesizing late proteins only after the viral genome has
been replicated? The relevance of this issue extends be- replicated, late protein synthesis does not compromise ge-
yond the DNA bacteriophages, since the regulated pattern nome replication. And it is less likely to compromise final
of T-even phage gene expression is characteristic of the virus assembly.
replication of DNA viruses in general. We next consider the mechanism by which T-even gene
A likely advantage to the virus of this pattern of regu- expression is regulated. We begin by noting that the pat-
lated gene expression is that it maximizes the efficiency of tern of T-even gene expression is characterized by the
viral protein synthesis. Consider that at the start of the following three features. First, phage DNA synthesis is
infection, when there may be only one viral genome in the dependent on previous phage-specific protein synthesis.
cell, the virus has relatively little capacity to generate large Second, early protein synthesis ceases at the time that
amounts of any particular protein. However, the virus DNA synthesis begins. Third, phage DNA synthesis is nec-
must make those virus-encoded proteins that are neces- essary for the switch from early to late gene expression to
sary for it to replicate its genome. Fortunately, these early occur. Thus, if phage DNA synthesis is blocked (e.g., by
proteins have an enzymatic role and thus are not needed ultraviolet [UV] irradiation of the phages or by muta-
in great amounts since they are used over and over again. tion), then early genes continue to be expressed and late
In contrast, the virus will need many copies of the differ- proteins are not made (see Fig. 2.4, in which the broken
ent proteins that make up the virus particle. Also, each of lines in the bottom part of the figure depict the pattern of

phage early protein synthesis in bacteria infected with that the superinfecting virus cannot express its late genes
conditional phage mutants that are unable to initiate unless it is permitted to replicate its own DNA. The actual
phage DNA synthesis). explanation is that replication introduces nicks into the
The experimental findings described above demon- DNA, which are necessary to activate the late transcrip-
strate that phage DNA synthesis is the event that triggers tional promoter. As then might be expected, the late pro-
the switch from early to late phage protein synthesis. How moter can be activated in vitro by introducing nearby
might phage DNA synthesis bring about this change in nicks.
the pattern of phage gene expression? We might hypoth- The role of DNA replication in the switch from early to
esize that the switch depends on phage DNA replication late protein synthesis is only a part of the story of the reg-
because DNA replication brings about an actual change in ulation of T-even phage gene expression. The story also
the phage DNA. An alternative hypothesis is that phage involves sigma () factors, which confer specificity on the
DNA synthesis might cause a change in cellular metabo- phage RNA polymerase. In the absence of an associated
lism that might then affect viral gene expression. factor, an RNA polymerase initiates transcription at ran-
Bruner and Cape distinguished between these alterna- dom points along a DNA template. But when an RNA
tive hypotheses in 1970. Since they used an experimental polymerase is associated with a factor, the polymerase
approach that exemplifies the clever molecular genetics of initiates transcription at those promoters that are specifi-
the day, we recount their experiment here. First, some cally recognized by that particular factor. The role of
necessary background that is probably already familiar to factors in transcription in fact was revealed by the analysis
most readers is provided. A single-base change in the DNA of the role they play in the regulation of T4 transcription.
can result in a mutant that is conditionally defective. Those phage genes that encode the DNA polymerase,
These conditional mutants are usually temperature sensi- the nuclease (which degrades host cell DNA), and the ac-
tive (ts). That is, they produce a protein that is not func- tivities responsible for nucleotide metabolism are all ex-
tional over as wide a range of temperatures as the wild-type pressed soon after infection. This is possible because the
protein. Generally, they are inactive at 39C and higher, cellular RNA polymerase, in association with a cellular
conditions under which wild-type proteins are still active. factor, 70, recognizes the phage promoters for these
Other single-base changes can result in the conversion of genes. Somewhat later, but still before phage DNA synthe-
a triplet that encodes an amino acid into one of the three sis, host RNA polymerase, still in association with 70,
triplets that cause chain termination. When such muta- but with the addition of the T4-specified positive regula-
tions occur in the coding region of a gene, they result in tory factor, motA, transcribes yet other phage genes. Tran-
premature chain termination, and the mutants are usually scription of the late genes requires, in addition to phage
nonconditionally defective. (They are not strictly non- DNA replication, the replacement of 70 with the phage-
conditional, since there are other mutations that can sup- specified factor, gp55. That switch of factors is medi-
press their effects.) They are often collectively referred to ated by yet another phage protein. Phage-specified regula-
as amber mutations, since the amber triplet UAG is one of tors of transcription and translation mediate the shutoff
the three chain-terminating triplets. Here is the actual ex- of the phages early genes. These regulators are among the
periment. First, cells are infected at 30C with a phage that late proteins that the virus encodes.
carries two mutations. One is a ts mutation (DNA-ts) in a The shutoff of host transcription is mediated by a
gene that is required for DNA synthesis. The second is an phage-induced modification of the host RNA polymerase
amber mutation in a late gene (am-late). At 30C, the core enzyme (i.e., the cellular core RNA polymerase is ad-
DNA-ts mutant protein is functional, and the double mu- enosine diphosphate [ADP]-ribosylated by the phage) a
tant replicates its DNA. However, it cannot express the few minutes after infection. One of the activities respon-
late function that contains the nonconditional amber le- sible for this modification is the protein gpalt, which the
sion. Consequently, no progeny phage can be produced. phage injects into the cell along with its DNA. The other is
The culture is then shifted to 42C and superinfected (i.e., gpmod, one of the phage proteins synthesized immedi-
infected with a second virus) with a virus that carries only ately after infection.
the DNA-ts lesion. This second virus cannot replicate its
DNA at the restrictive high temperature. However, it con- Phage Release: Lysozyme and the
tains a normal complement of late genes. Will it be able to rII Region
express its late genes, since they are normal, and since viral Next we consider how progeny phages manage to ensure
DNA replication has already occurred in the cells? Or will their release from the host cell at the end of their replica-
the superinfecting virus be unable to express its late genes, tion cycle, so that they are free to initiate another cycle of
since it has not replicated its own DNA? The answer is infection. Since bacterial cells are encased within rigid cell
Biosynthesis and Classification of Viruses 33

walls, it might be expected that phages require phage- conjunction with the replication of the cellular genome,
encoded functions to facilitate their release. This assump- and a copy of the phage genome is passed on to each
tion is correct. At late times in infection, T-even phages daughter cell at division.
express a lysozyme, which is a lytic enzyme that attacks In order for temperate bacteriophages to follow the ly-
the peptidoglycan layer in the bacterial cell wall, leading to sogenic course of infection, they must be able to restrict
cell lysis and liberation of the progeny phages. The phage the expression of their replicative functions. Thus, tem-
expresses another protein, encoded by its t gene, which perate phages make use of mechanisms that enable them
may disrupt the cytoplasmic membrane, in that way en- to express only a subset of their genes when in the ly-
abling the lysozyme to reach the peptidoglycan layer. sogenic state, in particular those genes which enable them
It is important that the timing of lysozyme expression to establish and maintain the lysogenic state.
is regulated, since if lysozyme were expressed too soon, Since lysogenic viruses express a small portion of their
then cell lysis might occur before infectious progeny genomes when in the lysogenic state, Lwoff was led to
phages were completely assembled (Fig. 2.4). The rII genes state that lysogens are the first example of a specific he-
of bacteriophage T4 are responsible for preventing pre- reditary property being conferred on an organism by a
mature expression of T4 lysozyme and consequent early specific extrinsic particle. In that regard, lysogens can be
cell lysis. The r in rII stands for rapid lysis, since that is medically relevant, since the lysogenic virus may express
the phenotype of rII mutants (Box2.5). deadly toxins, as in the case of Corynebacterium diphthe-
In contrast to the T-even phages, not all phages lyse riae and Clostridium botulinum, in which the toxins pro-
their bacterial hosts. The filamentous phages provide a duced by these potentially deadly bacteria are encoded by
particularly noteworthy example of such nonlytic phages. the lysogenic viral genomes that they carry.
(These phages have circular, single-stranded DNA ge- During the time that a bacteriophage is in the ly-
nomes that are enclosed within a 1-m-long tube com- sogenic state, the infection is said to be latent. Impor-
posed of about 2,000 copies of a single hydrophobic pro- tantly, when it might serve its purpose, a phage in the
tein.) Rather than lysing their bacterial hosts, progeny lysogenic state can activate and take up the course of
filamentous phages are continually extruded through the lytic growth. In so doing, the phage has the opportunity
cell envelope, while cell growth is affected only moder- to infect new hosts. The phage of E. coli is the para-
ately by the infection. (Cell growth is probably affected digm of a temperate phage. As we soon will see, the ability
because a portion of the cellular biosynthetic capacity is of phage to choose between the alternate existences
supporting replication of the phage.) Filamentous phage of lysogeny versus lytic replication depends upon a com-
growth can continue for long periods, making for a situa- plex regulatory mechanism of extraordinary elegance
tion intermediate between that of the lytic phages, such as (Box2.6).
the T-even phages, and temperate phages, such as phage, Before we consider the mechanism that determines
described below. Importantly, animal viruses too have whether phage will initiate a lytic versus a lysogenic
their counterparts to this range of interactions between course of infection, there are several other matters to at-
bacteriophages and their bacterial hosts. In the case of tend to. The first of these is the question of what advan-
animal viruses, these types of infections are referred to as tages phage might derive from having these alternative
acute, chronic, and latent. choices. The answer appears satisfyingly straightforward.
When phage infects healthy, metabolically active cells, it
chooses the lytic course. However, when phage infects
BACTERIOPHAGE l: LYSOGENY AND metabolically listless cells, it chooses the lysogenic course.
TRANSDUCTION Actually, the metabolic state of the host cell directly affects
In Chapter 1, we briefly introduced the temperate bacte- the viral switch that oversees the lytic versus lysogenic de-
riophages, noting that they do not necessarily replicate cision. Consequently, when conditions within a lysogen
in, and kill, their usual host cells. Instead, infection by improve, the genome can then activate and take up the
these viruses may follow a course in which a mostly qui- course of lytic growth. Moreover, if the cellular genome
escent phage genome is stably maintained in the host should be damaged by UV irradiation or by some other
cell; this virus-cell relationship is referred to as lysogeny means, the lysogenic virus can be activated to initiate lytic
(Box1.9). In some instances (e.g., bacteriophage ), the replication in order to find more fertile ground in a new
lysogenic phage genome is integrated into the cellular ge- cell. However, while in the lysogenic state, is able to re-
nome, whereas in other instances (e.g., phage P1), it is main within the protective environment of the host cell
maintained as an extrachromosomal plasmid. In either and replicate as a lysogen, concomitant with the replica-
case, the lysogenic phage genome is typically replicated in tion of the host cell.

Box 2.5
Adventures in the rII region To appreciate the immense significance of Benzers con-
tributions, we need to be reminded that classical genetics
In the late 1950s, Seymour Benzer carried out genetic made no distinction between genes as units that specified
experiments involving the rII genes of T4 phage, which a particular phenotypic trait, as units of mutation, and as
produced results of major importance to the newly emerg- units of recombination. Indeed, classical genetics envisioned
ing science of molecular biology. First, Benzers analysis of a gene as a single indivisible unit that embodied all three of
the T4 rII region provided the most detailed fine-structure these properties. Benzers experiments provided the distinc-
genetic map in existence. Second, he was the first to docu- tions between genetic units of function, of recombination,
ment genetic recombination between adjacent nucleotides. and of mutation, showing that a single pair of nucleotide
Third, he demonstrated that single nucleotides are the mini- bases is the unit of recombination and that a single nucle-
mal unit of mutation (muton). Fourth, and perhaps most otide is the minimal unit of mutation; and for the first time,
important, he devised the cis-trans test to define the unit of he gave the gene a clear operational definition. Moreover,
gene expression, the cistron. the cis-trans test could be used and indeed was used to
determine whether any two mutations are in the same or
The background for the last contribution is as follows. different functional genetic units.
When Benzer infected E. coli K cells with pairs of indepen-
dently isolated T4 rII mutants, he found that wild-type As a graduate student in physics during World War II,
T4 recombinants are generated at a very low frequency, Benzer participated in a project to develop better crystal rec-
indicating that all of the rII mutations are very close togeth- tifiers, a crucial component for radar. His research findings
er on the phage chromosome. In addition to finding rare on semiconductors later contributed to the development of
genetic recombinants between rII mutations, he also found the first transistors. When his interested shifted to biology,
that certain pairs of rII mutants actually replicated together he enrolled in Delbrucks summer phage course at the Cold
in E. coli K. That is, they complemented each other. Spring Harbor Laboratory, eventually joining Delbrucks
Phage Group at Caltech.
Next, Benzer found that all of the rII mutants could be
placed in either of two groups, designated A and B. All A From a human-interest standpoint, when Max Delbruck
mutants complemented all B mutants and vice versa. How- summoned Renato Dulbecco to his office to propose that
ever, mutants within the same group could not complement he, Dulbecco, work on animal viruses (Box1.11), Benzer
each other. Moreover, for complementation to occur, the was called inalong with him. Dulbecco relates, I im-
mutations also had to be on separate phage chromosomes; mediately expressed my interest, before Benzer could say
that is, they had to be in trans. Complementation did anything. Benzer, on the other hand, was not interested,
not occur if the mutations were on the same phage so everything was settled without delay. At that moment I
chromosome, that is, in cis. became an animal virologist. (See Dulbeccos chapter in
Phage and the Origins of Molecular Biology). Benzers passion
Together, these experimental results showed that whereas seems to have been the nature of the gene, and his contribu-
all of the rII mutations are very close together on the tions there were huge.
T4 chromosome, they nevertheless define two distinct
functions and that a mutation in either function results in In the 1960s, Benzers scientific interest shifted again, this
the mutant phenotype. But what might be the mechanistic time to the goal of developing a model system that might
explanation for complementation between rIIA and rIIB lead to insights into the genetic basis for behavior. He eventu-
mutants when expressed in cis, but not when expressed in ally settled on Drosophila melanogaster and founded the field
trans? The answer is that the rIIA and rIIB regions of the of neurogenetics. Benzer passed away in November 2007.
phage chromosome define separate units of transcription.
These units of transcription, or cistrons, are operationally Francis Crick in 1961 used the T4 rII region to make yet
defined by the cis-trans test. That is, mutations in separate another fundamental contribution to molecular biology,
cistrons complement each other when expressed in cis, but in this instance demonstrating that the genetic code is read
not when expressed in trans. Cistrons are thus the genetic three nucleotides at a time. In brief, he found that certain
units of function. Today, the term cistron is rarely used. T4mutant phages, which contained two rII mutations in cis,
Instead, the word gene has taken on the same meaning as displayed a wild-type-like phenotype. Specifically, some of
cistron. these mutations, referred to as + (plus), gave a wild-type
Box continues
Biosynthesis and Classification of Viruses 35

Box 2.5 (continued)

phenotype when in cis with other mutations, referred to as code was read as triplets, with each triplet specifying an
(minus). In contrast, + + and double mutants never amino acid, then a +++ or a triple mutant would also
displayed a wild-type phenotype. However, + + + and not change the reading frame downstream of the last mutation.
triple mutants could display a wild-type phenotype. The
explanation is that the + mutations are single-nucleotide The title at the head of this box is that of Benzers chapter in
insertions that change the reading frame by +1. Likewise, Phage and the Origins of Molecular Biology.
mutations are single-nucleotide deletions that change
the reading frame by 1. When a + mutation is paired with Cairns, J., G. S. Stent, and J. D. Watson (ed.). 2007. Phage and the
a mutation in cis, the reading frame downstream of the Origins of Molecular Biology: the Centennial Edition. Cold Spring
second mutation is restored, leaving the possibility that a Harbor Laboratory Press, Cold Spring Harbor, NY.
functional protein might yet be translated. Similarly, if the

Box 2.6
Although phage resembles the T-even phages in key lysogen, as later discovered by Esther Lederberg in 1951.
ways (e.g., the head is filled with DNA and it is attached Esther Lederberg was the wife of Joshua Lederberg, who
to a slender tail, through which it injects its DNA into was awarded the Nobel Prize in 1958 for discovering genetic
susceptible host cells), Delbruck and his group of phage recombination associated with sexual conjugation in bacte-
researchers were initially not interested in lysogeny, primar- ria. The couple worked together to develop the technique of
ily because the T-even phages they had been studying did replica plating in 1952, a technique that enabled the selection
not display this phenomenon. They also may have been of bacterial mutants from among hundreds of bacterial colo-
influenced by the view of dHerelle, one of the discoverers nies on a plate and, more importantly perhaps, proof of the
of bacteriophages (Chapter 1), who believed that infection spontaneous origin of mutants with adaptive advantages.
by a bacteriophage inexorably leads to cell death. dHerelle
thought that so-called lysogens were merely contaminated Esther Lederberg accidentally isolated nonlysogenic or
with a bacteriophage. cured derivatives of E. coli K-12 that could be infected
by samples of culture fluid from the parental K-12strain,
It also is historically interesting that the original laboratory which was sporadically producing low levels of phage. As
strain of E. coli (i.e., E. coli K-12), which was isolated in explained in the text, bacterial strains that carry a lysogenic
1922 from a patient with an intestinal disorder, was a virus are immune to superinfection by the same virus.

Next, note that the genome is linear in the virus par- need to consider the elements of the critical control re-
ticle but that it circularizes upon injection into the cell, gion, which is located in the right-most third of the linear
doing so via cohesive sites at its ends. The ability of the genome (Fig. 2.6). The control region is referred to as
DNA to circularize is a key factor that enables it to inte- the immunity region. It encodes the diffusible CI pro-
grate into the cellular genome and thus be maintained in tein, which is the key repressor that suppresses lytic phage
its lysogenic state. During lytic growth, the circular ge- growth. The immunity region derives its name from the
nome is replicated by means of a rolling-circle mecha- fact that CI also suppresses the growth of any superinfect-
nism, which is discussed in detail in Chapter 14. For the ing phages, as well as of the resident integrated prophage.
sake of comparison, the genome is considerably smaller The suppressing effect of CI is specific to phage. CI
than the genomes of the T-even phages, encoding about has no effect on infection by other bacteriophages. The
60 proteins, versus the 160 or so proteins encoded by the immunity region also contains the binding sites for the CI
larger viruses. protein, which are known as operators, since they are
Before we can understand how phage decides analogous to the operators that control transcription of
whether infection will be lytic versus lysogenic, we also the lac operon of E. coli.

int xis cIII N cI
cro cII

cro mRNA

Figure2.6 (A) The control region, referred to as the immunity region, is responsible for the establishment and
maintenance of lysogeny. It encodes the diffusible CI protein, which is the key repressor that suppresses lytic
phage growth. The immunity region also contains the binding sites for the CI protein, designated OL and OR, which
are located on each side of the cI gene. The genome also contains two main promoters where the E. coli RNA
polymerase can initiate transcription. Each of these promoters is located in the immunity region, one on each side
of the cI gene. They are designated PL and PR and overlap OL and OR, respectively. OL initiates transcription leftward,
and OR initiates transcription rightward. Another promoter, designated PRM (RM stands for repressor mainte-
nance), is located near PR but initiates transcription in the opposite direction, to generate the mRNA that encodes
the CI protein. Pint is the promoter for the int gene, which encodes the integrase that catalyzes the site-specific re-
combination between the phage DNA and the cell DNA. (B) OL and OR are each comprised of three segments, OR1,
OR2, and OR3in the case of OR, and each of these segments binds a dimer of CI. The figure shows the topological
relationship between the elements of OR and PRM and PR. When CI binds to OL and OR, it inactivates PL and PR, re-
spectively, while activating PRM, thereby ensuring continued synthesis of CI while the phage is in the lysogenic state.
When phage DNA first enters a cell, early viral transcription can begin from PL and PR, resulting in the produc-
tion of the N and Cro proteins. N is a transcriptional antiterminator that enables the cellular RNA polymerase to
continue transcription beyond the ends of the N and cro genes, so that the cII gene, among other early genes, might
be transcribed. Cro is a repressor, which, like the CI protein, binds to OL and OR. But whereas CI has a preference
for OR1, Cro has a preference for OR3. PRM is blocked when Cro is bound to OR3, but not when CI is bound to OR1
and OR2. CI dimers bind cooperatively, such that when a CI dimer binds to OR1, it facilitates the binding of a sec-
ond dimer to OR2. In contrast to CI, Cro does not bind cooperatively. Thus, CI can repress expression of PL and PR,
thereby blocking lytic infection, whereas Cro can repress expression of PRM, and thus of cI, thereby enabling lytic in-
fection. PRE is a second promoter from which transcription of cI can be initiated. The decision as to whether infec-
tion will be lytic or lysogenic actually depends on the amount of CI that might be present as a result of transcrip-
tion initiated at PRE. If there is not enough CI present to prevent transcription from PL and PR, then the outcome
will be lytic infection. But if sufficient CI is produced from PRE to repress PL and PR, then production of early pro-
teins will be blocked, including that of Cro. Moreover, CI produced as a result of transcription from PRE can activate
further expression of cI from PRM, thereby facilitating the establishment and maintenance of lysogeny. The level of
CI generated from PRE is determined by the level of CII, which positively regulates initial transcription of cI by
binding to PRE, thereby activating the otherwise weak PRE. The level of CII in turn depends on the metabolic state
of the cell. CII is inactivated by cellular proteases that are more abundant in cells that are actively replicating.

The genome contains two main promoters where the comprised of three distinct segments (e.g., OL1, OL2, and
E. coli RNA polymerase can initiate transcription. Each of OL3), and each of these segments binds a dimer of CI.
these promoters is located in the immunity region, one on When phage initially injects its DNA into a cell, and
each side of the cI gene. They are designated Pleft (PL) and if there is no prophage present that might be synthesizing
Pright (PR), and they initiate transcription leftward and the CI immunity repressor, then immediate early viral
rightward, respectively, from cI. Another promoter, desig- transcription begins from PL and PR, resulting in the pro-
nated PRM (RM stands for repressor maintenance), is lo- duction of the N and Cro proteins (Fig. 2.6). The N pro-
cated near PR but initiates transcription in the opposite tein is a transcriptional antiterminator that enables the
direction, to generate the mRNA that encodes the CI cellular RNA polymerase to continue transcription be-
protein. yond the ends of the N and cro genes, into the adjacent
The binding sites in the immunity region for the CI viral early genes. N acts by a mechanism analogous to that
protein are designated OL and OR, since they are analo- of the human immunodeficiency virus (HIV) Tat protein
gous to operators on bacterial genomes. OL overlaps PL, (see Chapter 21). In brief, N binds to a sequence called
and OR overlaps PR (Fig. 2.6). Each operator actually is NUT (which stands for N utilization), which is present on
Biosynthesis and Classification of Viruses 37

nascent transcripts that are initiated at PL and PR. NUT The level of CI generated from PRE is determined by
positions the bound N protein so that it is able to interact the level of CII. (Recall that cII is located just down-
with cellular factors that enhance the processivity of the stream of cro and is transcribed from PR [Fig. 2.6].) CII
cellular RNA polymerase. Thus, when N is present, it en- positively regulates initial transcription of cI by binding
ables the polymerase to transcribe the cII gene, immedi- to PRE, thereby activating the otherwise weak PRE. So, the
ately adjacent to cro. As soon will be seen, cII plays a cru- lytic replication versus lysogeny decision depends on the
cial role in determining whether the infection will be lytic level of CII. Now here is a really neat part of the story.
or lysogenic. Moreover, N enables the polymerase to con- The level of CII in turn depends on the metabolic state
tinue transcribing leftward and rightward (past cro), to of the cell. CII is inactivated by cellular proteases that
express the remaining early genes, including the late gene are more abundant in cells that are actively replicating.
activator. Thus, infection tends to be lytic in exponentially repli-
The Cro protein is a repressor, which, like the CI pro- cating cells and lysogenic in stationary-phase cell cul-
tein, binds to OL and OR. However, recall that the operators tures. The regulatory machinery that maintains the ly-
are comprised of three distinct segments (Fig. 2.6; look sogenic state is indeed very effective, since lysogeny is as
carefully at the topological relationships of PRM and PR stable a characteristic of a bacterial strain as any other
with OR3, OR2, and OR1.). Whereas CI has a preference for inherited characteristic.
OR1, Cro has a preference for OR3. Importantly, PRM is The above account of the lysis-versus-lysogeny deci-
blocked when Cro is bound to OR3, but not when CI is sion is somewhat watered down, since it does not consider
bound to OR1 and OR2. Apropos that, CI dimers bind co- items such as the interaction between CI dimers in the
operatively, such that when a CI dimer binds to OR1, it regulatory region and the role of CIII in inhibiting the
facilitates the binding of a second dimer to OR2. In con- cellular proteases that inactivate CII. Nevertheless, I hope
trast to CI, Cro does not bind cooperatively. Thus, CI can that it succeeds in conveying the elegance of the genetic
repress expression of PL and PR, thereby blocking lytic switch. For those whose interest may have been piqued by
infection, whereas Cro can repress expression of PRM, and the above, I strongly suggest a marvelous text by Mark
thus of cI, thereby enabling lytic infection. Ptashne (2004), who contributed much to the under-
The plot thickens somewhat because the cI promoter, standing of the lysis-versus-lysogeny decision.
PRM, is a weak promoter that actually is activated when CI Next, we briefly consider the means by which the
binds to OR2 (Fig. 2.6). CI has this effect on PRM by directly circularized genome is integrated into the cellular
facilitating binding of the RNA polymerase to PRM. Thus, chromosome, a process that perpetuates the phage genome
CI activates its own gene while suppressing transcription in the cell and its descendants, until an opportune moment
from PR. This raises the following question. If CI is neces- might arise when the viral genome might be activated to
sary to activate its own gene, then where does the CI come undergo lytic replication. In addition, we consider the
from to do this? The answer is crucial, since CI is neces- means by which the lysogenic phage can resume productive
sary to establish and maintain the lysogenic state. Here is infection, when it becomes advantageous for it to do so.
the explanation. Concerning integration, when phage makes the de-
The control region has yet another promoter, desig- cision to lysogenize, it expresses a site-specific recombi-
nated PRE (RE stands for repression establishment), from nation system that recombines the attP site on the circu-
which transcription of cI can be initiated (Fig. 2.6). PRE is larized phage genome with the attB site on the E. coli
located just downstream of cro, so that transcription from chromosome (located between the galactose [gal] and
PRE is in the opposite direction (and from a different DNA biotin [bio] operons). By recombining a circularized phage
strand) from transcription from PR. Now, the decision as genome with the bacterial chromosome, a linear viral ge-
to whether infection will be lytic or lysogenic depends nome is integrated into the bacterial chromosome.
upon the amount of CI that might be present as a result of The attP site is about 250 base pairs (bp) in length,
transcription initiated at PRE. If there is not enough CI to whereas attB is about 25 bp long. Each of these sequences
prevent transcription from PL and PR, then the outcome contains a common 15-bp-long sequence, at which the
will be lytic infection. But if sufficient CI is produced from integration of the chromosome takes place. Neverthe-
PRE to repress PL and PR, then production of early proteins less, the site specificity of this crossover reaction is largely
will be blocked, including that of Cro. Moreover, CI pro- due to the phage Int protein, which recognizes the attP
duced as a result of transcription from PRE can activate and attB sites on the phage and bacterial chromosomes,
further expression of cI from PRM, thereby facilitating the respectively, and catalyzes the integration event.
establishment and maintenance of lysogeny. So, the cru- What might control the expression of int? Congratula-
cial question is, what determines the critical level of CI? tions if you correctly surmised that in addition to positively

controlling PRE, CII also positively regulates Pint, the weak is a specialized transducing phage. However, other bacte-
promoter for int (Fig. 2.6). A very neat picture! riophages, such as phage P1 of E. coli and P22 phage of
Integration and lysogeny have a flip side: induction of Salmonella enterica serovar Typhimurium, can pick up
the lysogenic viral genome and the excision reaction that any gene segment from the host chromosome, a phenom-
frees it from its integrated state. As noted above, the rela- enon referred to as generalized transduction. (Experi-
tive levels of CI and Cro, which determine whether the ments with phage P22, carried out by Norton Zinder and
infection will be lysogenic or lytic, are affected by the met- Joshua Lederberg in 1952, led to the discovery of trans-
abolic conditions of the host cell. Thus, under conditions duction.) In these instances, transduction results from the
of robust cell growth, the regulatory circuit may favor in- headfull packaging mechanism by which the phage
duction of the lysogenic phage and its vegetative replica- heads package DNA. Each phage head is normally filled
tion. The regulatory circuits may also occasionally fail with a little more than one phage length of DNA from a
spontaneously, so that there usually are a few cells in a concatemer of progeny phage genomes, which have been
lysogenic culture that are producing viruses at any mo- generated by rolling-circle replication. Although these
ment. Apropos the excision process, another early gene, phages typically recognize their own DNA as appropriate
xis (Fig. 2.6), encodes the excisionase, Xis, which together for packaging, they occasionally incorrectly package host
with Int carries out the reverse process of excision, thereby DNA, presumably because they mistake a sequence in the
freeing a circularized genome from the bacterial chro- cellular DNA for the special phage packaging signal. Thus,
mosome. The mRNA that encodes both Int and Xis is these transducing phage particles carry cellular DNA se-
transcribed from the reactivated PL promoter. quences only. Also in contrast to l phage, the P1 and P22
Before leaving the topic of lysogeny, note the follow- genomes are maintained as extrachromosomal plasmids.
ing two points regarding the induction of a lysogen and The cellular DNA carried by a generalized transducing
its excision from the cellular genome. First, induction of phage is likely to be linear due to an absence of cohesive
the lysogen also occurs under conditions in which the ends that might facilitate its circularization. Thus, for the
host cell has been damaged, perhaps irreversibly. Induc- transduced DNA to integrate so that it might be stably
tion under these circumstances may enable the phage to inherited, it must undergo two homologous recombination
find a new healthy host cell in which it might reestablish events with the recipient host chromosome, catalyzed in
itself. But how might the phage be so clever as to know this instance by the host recombination enzymes. One
to reactivate in the traumatized cell? The answer is that consequence is that a portion of the recipient chromo-
when the cell undergoes extensive damage, such as from some is replaced by a portion of the incoming DNA. De-
UV irradiation, it activates its so-called SOS repair sys- spite the requirement for two homologous recombination
tem. To do so, the cell produces a particular protease, events, there still is about a 10% probability that the in-
which attacks and destroys the repressor of the host SOS coming chromosomal DNA will be recombined into the
genes. So, our clever phage has evolved its CI protein so recipients chromosome.
that it too is inactivated by that host protease, and in that Generalized transducing phages played a major role in
way the lysogen is activated in the traumatized cell. the analysis of several bacterial genomes. What sorts of
The second point is that whereas excision is generally a information might they have provided? Importantly, the
precise reversal of the site-specific integration reaction, distance between two genes on the bacterial chromosome
nevertheless, it occasionally goes awry, so that the released can be inferred by the frequency at which they are cotrans-
phage genome may contain a portion of the bacterial duced. Thus, it was possible to use generalized transduc-
chromosome. Since the genome specifically inserts be- ing phages to generate fine-structure maps of small re-
tween host genes involved in galactose metabolism and gions of bacterial chromosomes.
biotin synthesis, an incorrectly excised genome may Interestingly, whereas phage inserts its DNA into the
contain a portion of one of these host genome sequences. host cell chromosome at a specific site, Mu phage (not
Some of these incorrectly excised phage genomes may yet previously mentioned) inserts its DNA anywhere in the
be able to replicate and even integrate into the genome of host chromosome. (Although Mu may incorporate host
a new host cell. In that case, when in the lysogenic state in sequences into its DNA when it is excised, it is not a trans-
the new host cell, they may be able to express the captured ducing phage, because when Mu infects a new host, the
host cell gene sequence. This phenomenon is referred to Mu DNA, but not the variable sequences of DNA acquired
as transduction. from the previous host, is incorporated into the recipi-
Since the genome inserts at only one specific site in ents chromosome by the Mu integration process.)
the bacterial chromosome, it only transduces cellular gene As we will see in subsequent chapters, many animal
sequences adjacent to its integration site. That is, phage viruses too establish long-term associations with their
Biosynthesis and Classification of Viruses 39

eukaryotic hosts, which may have serious clinical conse- Also, since (i) bacteria are haploid and (ii) phages are
quences. Examples include HIV, hepatitis B virus, the capable of transducing host cell genes, phages make an
papillomaviruses, and herpesviruses. Each of these differ- important contribution to the genetic diversity of bacte-
ent animal viruses has evolved a different means for re- ria. Moreover, since temperate phages often encode bacte-
maining associated with the individuals it has infected, rial virulence factors (e.g., toxins) and antibiotic resistance
and each of these means is different from that evolved by genes, phages play important roles in particular instances
bacteriophage ; this is yet another example of different of bacterial pathogenesis and contribute to the general
viruses evolving different strategies to achieve the same problem of antibiotic resistance.
ends. Nevertheless, bacteriophage is able to select Finally, despite the early unsuccessful attempts by
between two alternative existences, lytic growth versus dHerelle to use bacteriophages as therapeutic agents in
latency, choosing the more suitable of these outcomes in the treatment of bacterial diseases (see Chapter 1), this
light of the health and growth state of the host, a feat approach has reemerged in new guises, largely because of
that is beyond the capabilities of any known animal the growing problem of antibiotic resistance among bac-
virus. terial pathogens. For example, the recently isolated lytic
KMV phage of Pseudomonas aeruginosa is highly viru-
lent against many clinical isolates of its host cell. That fact
SOME FINAL COMMENTS ON and the short replication time of KMV make the phage
BACTERIOPHAGES or a phage product (e.g., the phage lysozyme) potentially
This chapter began with the continuing account of how attractive as therapeutic agents against P. aeruginosa.
analysis of the T-even phages revealed key general features
of virus replication. In addition, we also considered some
aspects of T-even phage replication that are specific for INTRODUCTION TO THE ANIMAL
these viruses, as well as features of bacteriophage that VIRUSES
are relevant to the phenomenon of lysogeny. Still, despite We begin our consideration of the animal viruses with a
the fact that many years have passed since bacteriophages brief introduction to virus structure and follow with a
served as models for animal viruses, much more could be short account of how animal viruses invade their host
said about bacteriophages. Other bacteriophages, such cells. These topics are covered in more detail in subse-
as the small phages with single-stranded DNA genomes quent chapters, in the context of the individual virus
(in particular, X174) and the filamentous DNA phages families.
(the prototype is fd), also played important roles in the
development of virology. Animal Virus Structure
The single-stranded RNA phages (e.g., Q and MS2) As noted earlier, viruses are assembled from multiple cop-
also were not discussed above. However, these phages are ies of one or more kinds of identical protein subunits.
considered in detail in the chapter on the Picornaviridae, a This is partly, but not entirely, because viruses are limited
family of small RNA viruses that includes the rhinovi- in the number of genes that they might be able to package
ruses, which are responsible for many of our common within their capsids. Indeed, some viruses must make do
colds, and poliovirus. The reason for including the RNA with as few as five genes.
phages in that context is to compare their genome organi- The protein building blocks or subunits from which
zation and patterns of gene expression with those of the viruses are assembled are arranged to generate either of
eukaryotic picornaviruses. two general types of capsid structures. One type displays
Before leaving the bacteriophages, we should be aware helical symmetry. The other type displays icosahedral
that there has been a recent resurgence of interest in them symmetry. In order to generate symmetry of either type,
for their own sake. One reason is that new genomics-based each of the identical subunits makes the same sorts of con-
studies revealed a more pervasive relationship between tacts with its neighboring subunits. This is an important
these parasites and their prokaryotic host cells than was point that helps to account for the extraordinary tensile
previously imagined. Consequently, since bacteria are es- strength that is characteristic of virus capsids. If the con-
timated to constitute more than one-half of the Earths tacts between subunits are the same, then they can be opti-
biomass, and since phages are widespread among the bac- mized to be most favorable inall instances. But if they are
teria, virulent phages may profoundly affect more than not the same, then some contacts will be weaker than oth-
one-half of the Earths biomass. Indeed, it is estimated ers. Thus, another reason for viruses to build their capsids
that every 2days, one-half of the bacteria on the planet from identical subunits is that it enables them to assemble
are killed by bacteriophages! capsids of extraordinary tensile strength. Remarkably, in

NP subunit
() RNA = 0.41 nm


1,000 nm

Sendai virus

Coat protein
subunit = 0.14 nm
(+) RNA

18 nm (+) RNA

300 nm Coat protein subunit

Tobacco mosaic virus

N and M protein subunits

() RNA = 5 nm

27 nm 75 nm

180 nm

Vesicular stomatitis virus

Figure2.7 Examples of virus structures with helical symmetry. Tobacco mosaic virus (TMV) is a prototypical single-
stranded RNA virus. Its RNA genome is of the plus (+) sense (see the text). A single copy of the TMV RNA genome is en-
closed within a helical structure comprised of a single viral protein. There are 161/3 copies of that protein per turn of the
helix; each of these protein subunits makes the same interactions with its neighbors, and each interacts with three nucle-
otides in the RNA. Sendai virus (a paramyxovirus) has a single-stranded RNA genome of the minus () sense (see the text).
Its structure resembles that of TMV, but whereas the TMV helix is right-handed, the Sendai virus helix is left-handed.
(Viruses having minus-sense RNA genomes generally contain nucleocapsids that possess helical rather than icosahedral
symmetry. Moreover, the helical nucleocapsids of minus-sense RNA viruses are enclosed within lipid envelopes [not shown
here, but see Fig. 2.11]. Also, the genomes of these viruses are associated with multiple copies of the viral RNA polymerase
molecules, in addition to the NP protein, which serves to encapsidate the RNA.) Vesicular stomatitis virus (a rhabdovirus),
like Sendai virus, contains a minus-sense single-stranded RNA genome. Notice that the Sendai virus helical capsid contains
about 30 turns of uniform diameter, followed by 5 or 6 turns of decreasing diameter, thereby conferring on these viruses
their characteristic bullet shape. Adapted from S. J. Flint, L. W. Enquist, R. M. Krug, V. R. Racaniello, and A. M. Skalka, Princi-
ples of Virology: Molecular Biology, Pathogenesis, and Control, 2nd ed. (ASM Press, Washington, DC, 2003), with permission.

most, but not all instances, the protein subunits interact icosahedra have 12 vertices, each of which is a point at
via noncovalent contacts. which five of the triangular faces come together. The sim-
Viruses displaying helical symmetry enclose their ge- plest icosahedral viruses contain 3 protein subunits per
nomes within long tubes that are comprised of helically face and thus 60subunits in the entire capsid. In point of
arranged protein subunits (Fig. 2.7). The length of the fact, an icosahedron containing 60subunits is the simplest
tube reflects the length of the particular viral genome. structure, containing the least number of subunits, which
An icosahedron is a regular 20-sided polygon, where can form a shell that completely encloses a space of maxi-
each face is an equilateral triangle (Fig. 2.8). Note that mal volume. Interestingly, the icosahedral structure is also
Biosynthesis and Classification of Viruses 41

Structural unit



Figure2.8 The building plan of viruses that have icosahedral symmetry. An icosahedron is a regular 20-sided poly-
gon, where each face is an equilateral triangle. Note that icosahedra have 12 vertices, located at points where five of the
triangular faces come together. (A) The simplest icosahedral viruses contain 3 protein subunits per face and thus
60subunits in the entire capsid. (B) Icosahedral capsids are assembled from pentamers of the individual subunits,
which become the vertices of the icosahedron. The five subunits that form the vertices are connected by lines in the
diagram. (C) Making a larger icosahedral capsid requires adding more protein subunits, in accord with strict rules
for how this may be done. Essentially, each of the 20 triangular faces of the simplest icosahedron may be divided into
smaller triangles, each of which contains three subunits. Thus, icosahedral capsids always contain a number of sub-
units that is a multiple of 60. The resulting structure retains the fivefold, threefold, and twofold axes of symmetry that
are characteristic of icosahedra. Looking carefully at the arrangements of subunits in the larger icosahedral capsids,
notice that whereas the subunits at the 12 icosahedral vertices are arranged as pentamers, all other subunits are
arranged in groups of 6 (i.e., as hexamers). In some instances, the hexamers or hexons may be comprised of the same
protein subunits that constitute the pentons at the vertices. In other instances they may be comprised of a different
protein. (D) In the larger icosahedral virus, the pentons at the vertices are surrounded by five hexons. Each of the
hexons is surrounded here by four hexons and two pentons. Thus, notice that the pentons are foci of fivefold rota-
tional symmetry, and the hexons are foci of twofold rotational symmetry. (E) Interactions between subunits in
adjacent pentamers and hexamers cannot all be the same. Yet they can be very similar, according to the principle of
quasiequivalent bonding, in which each of the subunits can be bound to its neighbors in similar, but not exactly
identical, ways (e.g., the head-to-head interactions are very similar throughout the particle, regardless of whether
they are between pentamers and hexamers or between hexamers and hexamers). Panels A to D were adapted from
J. A. Levy, H. Fraenkel-Conrat, and R. A. Owens, Virology, 3rd ed. (Prentice Hall, Upper Saddle River, NJ, 1994), and
panel E was adapted from B. N. Fields, D. M. Knipe, and P. M. Howley (ed.), Fundamental Virology, 3rd ed.
(Lippincott-Raven, Philadelphia, PA, 1996), with permission.

the basis for the geodesic dome, which is exceptionally look like an icosahedron per se; however, it always retains
rigid, and which is said to require the least energy to as- the fivefold, threefold, and twofold axes of symmetry that
semble (Box2.7). are characteristic of icosahedra (see if you can identify
Icosahedral capsids are actually assembled from pen- these axes of symmetry in Fig. 2.8).
tamers of the individual subunits, which become the ver- Looking carefully at the arrangements of subunits in
tices of the icosahedron. That is, the capsids are assembled the larger icosahedral capsids, notice that the subunits at
from morphological units that correspond to their verti- each of the 12 icosahedral vertices are arranged in groups
ces, rather than their faces. In the simplest cases, each of of 5. However, all of the remaining subunits are arranged
the identical triangular faces contains three identical sub- in groups of six, that is, as hexamers. In some instances
units, as noted above. (e.g., many bacteriophages), these hexamers or hexons
As in the case of helical viruses, making a larger icosa- may be comprised of the same protein subunits that con-
hedral capsid requires adding more protein subunits. stitute the pentons at the vertices. In other instances (e.g.,
However, whereas helical viruses can be made larger by the adenoviruses), they may be comprised of a different
simply extending the length of the helix, making a larger protein. In cases where pentons and hexons are comprised
icosahedral virus is somewhat more complicated. Casper of the same protein subunit, the interactions between
and Klug discovered the strict rules that determine how subunits in the pentons cannot be exactly identical to
this can be done. Essentially, each of the 20 triangular their interactions in the hexons. Yet they can be very simi-
faces of the icosahedron may be divided into smaller tri- lar. Moreover, the interactions between subunits in adja-
angles, each of which contains three subunits. Thus, icosa- cent pentamers and hexamers cannot all be the same. Yet
hedral capsids always contain a number of subunits that is they too can be very similar. For these reasons, Casper and
a multiple of 60 (Fig. 2.8). The structure may no longer Klug introduced the concept of quasiequivalent bonding,

Box 2.7 0.1

Buckminster Fuller (18951983) was an American

architect and an early environmentalist, best known for
designing the geodesic dome and for popularizing the
term Spaceship Earth. More recently, carbon molecules,
known as buckyballs, or fullerenes, were named for their
resemblance to Fullers domes.

The geodesic dome is essentially an icosahedron, with

its faces subdivided into equilateral triangles. However,
the vertices of the geodesic dome are projected onto the
surface of a circumscribing sphere. Regardless, the icosa- A B
hedral building plan provides the same advantages to the Figure2.9 (A) Animal viruses with helical symmetry are generally
architect as it does to viruses. First, there is its inherent sta- enclosed within lipid envelopes. The electron micrograph shows a
bility, which results from the rigidity of triangles. Second, a filamentous form of simian virus 5 (a paramyxovirus) budding from
sphere, which is more closely approached by an icosahedron the surface of an infected cell. (B) Some viruses that have icosahedrally
than by any other polygon, has the least surface area for a symmetric capsids also bud through a particular cellular membrane,
thereby acquiring envelopes. The electron micrograph shows a
given volume. This is important to the architect, since it budding type C retrovirus. Note the continuity of the viral envelope
minimizes construction costs and heat loss as well. Thus, with the cellular plasma membrane. Electron micrographs by R. W.
the geodesic dome and the icosahedral viral capsid each Compans and P. W. Chopin (A) and C. Moore (B); reprinted from
benefit from the advantageous shape of the icosahedron. S. E. Luria, J. J. Darnell, D. Baltimore, and A. Campbell, General
And, in each instance, by subdividing the triangles, ac- Virology, 3rd ed. (John Wiley & Sons, New York, NY, 1978), with
cording to certain rules, the structure comes closer to an
actual sphere, while maintaining its stability.
Depending on the particular virus, that membrane might
Geodesic domes were being constructed for at least a de-
be the plasma membrane, the nuclear membrane, the en-
cade before electron microscopy revealed the icosahedral
doplasmic reticulum membrane, or a Golgi membrane.
nature of virus shells. But bear in mind that viruses came
up with the idea well before Buckminster Fuller did.
Entry of Animal Viruses
All virus infections begin when the virus attaches to a spe-
cific receptor on the surface of a susceptible or permissive
host cell. (A susceptible cell is sometimes taken to mean
whereby each of the subunits can be bound to its neigh- one that expresses a receptor that the particular virus can
bors in similar, but not exactly identical, ways (e.g., the bind to. A permissive cell additionally expresses whatever
head-to-head interactions are very similar throughout the other factors might be required to support productive
particle, regardless of whether they are between pentam- replication of the virus.) The particular receptor used by
ers and hexamers or between hexamers and hexamers an animal virus largely determines its tissue tropism (i.e.,
[Fig. 2.8]). Importantly, in most viruses, all subunits in- choice of cell type) within the animal host and, conse-
teract via noncovalent bonds and, therefore, are arranged quently, the pattern of pathology that might result from
to make maximal contact. the infection. HIV, which causes acquired immunodefi-
Note that even in the cases of the smallest icosahedral ciency syndrome (AIDS), provides an excellent example
viruses, which contain 60subunits in the capsid, the sub- of this point (Chapter 21). HIV uses the helper T-cell pro-
units need not be identical. For example, the poliovirus tein CD4 (Chapter 4) for its receptor. This explains why
capsid shell is comprised of three different kinds of protein HIV targets helper T cells and thus causes AIDS, which is
subunits. Other variations of the above general themes are the severest form of immunodeficiency known.
discussed in the context of the individual virus families. After binding to their receptors, animal viruses, like
Animal viruses with helical symmetry are generally bacteriophages, must transport their genomes into the
enclosed within lipid envelopes. In addition, some icosa- host cell for replication to begin. However, their means for
hedral viruses also are enclosed within lipid envelopes. En- doing so is in marked contrast to the corresponding step
veloped viruses generally acquire their envelopes when they in a bacteriophage infection. Phages inject their genomes
bud through a particular cellular membrane (Fig. 2.9). into their bacterial hosts. They must do so in order to
Biosynthesis and Classification of Viruses 43

Stage 1 Stage 2
Extracellular macromolecule (ligand)


Stage 3 Stage 4

coated pit

Budding 0.2 m

Clathrin-coated vesicle

Figure2.10 Approximately two-thirds of animal viruses are estimated to hijack the cellular receptor-mediated
endocytosis pathway, which the cell normally uses for the uptake of specific macromolecules. (A) Receptor-mediated
endocytosis has the following characteristics. Ligands bind to their specific receptors, which diffuse in the plane of the
membrane into clathrin-coated pits. These coated pits continuously and constitutively mature into coated vesicles that
pinch off from the plasma membrane, thereby delivering their cargo into the cell. The endocytic vesicle then traffics
to, and fuses with, an early endosome, into which it releases its cargo. Note that early endosomes are somewhat acidic
because of the activity of a vacuolar proton pump. Some receptors traffic back to the plasma membrane via recycling
endosomes that bud from the early endosome, whereas others traffic to late endosomes, which then fuse with, or ma-
ture into, lysosomes. (B) The electron micrographs show four sequential stages from the development of a clathrin-
coated pit to the release of a clathrin-coated vesicle. Adapted from M. M. Perry and A. B. Gilbert, J. Cell Sci. 39:
257272, 1979, with permission. (C) Semliki Forest virus (a togavirus) is seen associated with a coated membrane
invagination (left) and enclosed within a coated vesicle (right). Electron micrographs by A. Helenius, reprinted from
K. Powell, J. Cell Biol. 170:1018, 2005.

breach the cell walls that are characteristic of prokaryotic with clathrin-coated pits in the plasma membrane. These
cells. Since animal cells do not have cell walls, many ani- coated pits continuously and constitutively mature into
mal viruses are able to initiate entry by subverting the very coated vesicles that pinch off from the plasma membrane,
mechanisms that their eukaryotic host cells use for their thereby delivering their cargo into the cell. However, bear
normal ingestive processes. in mind that the endocytosed virus, as well as any other
Most animal viruses hijack the cellular receptor- endocytosed cargo, is still contained within the endocytic
mediated endocytosis pathway to mediate their entry. vesicle. Thus, the virus still must breach at least one cel-
The cell normally uses this process to specifically inter- lular membrane in order to initiate its replication.
nalize particular macromolecules (Fig. 2.10). Receptor- The endocytic vesicle generally traffics to, and fuses
mediated endocytosis has the following characteristics. with, an endosomal compartment, thereby releasing its
Ligands bind to their specific receptors, which associate cargo into the endosomal lumen. The mildly acidic pH of

the endosome then induces the virus particle to undergo

structural rearrangements that can promote release of the
Box 2.8
virus or its genome into the cytosol. Both enveloped and
The mechanism by which enveloped viruses acquire their
nonenveloped viruses may access the cytosol in this way.
spike proteins begins by essentially the same mechanism
However, we have not actually explained the processes by that targets cellular membrane proteins to specific host
which animal viruses breach a cellular membrane. Breach- cell membranes. This process is described in detail in
ing a biological membrane is by no means a trivial event, Chapter 8. In brief, it occurs as follows. Integral mem-
since membranes by design serve as barriers and viruses brane proteins contain so-called signal sequences. As
are very large relative to other entities that necessarily do these signal sequences emerge on the ribosome-bound
cross membranes. The means by which animal viruses nascent polypeptide, they target the nascent polypep-
cross cellular membranes are better understood for some tide and ribosome to translocation complexes located at
viruses than for others. One well-studied case is that of the endoplasmic reticulum (ER) membrane. The signal
influenza virus (Chapter 12), and another is that of HIV sequence then inserts into channels in the ER membrane.
(Chapter 21). Some general points regarding how viruses As translation of the nascent protein continues, the
cross cellular membranes are as follows. protein moves through the channel. What then causes
Enveloped viruses contain virus-encoded spike pro- the protein to remain as an integral membrane protein?
teins, which are embedded in their envelopes (Box2.8). Like cellular integral membrane proteins, virus spike
The low pH of the endosomal compartment may trigger proteins contain a hydrophobic sequence that causes
conformational changes in these spike proteins, which has them to remain embedded as transmembrane proteins
the effect of exposing hydrophobic regions on them. These in the ER membrane. How then does the virus acquire its
regions, known as fusion peptides, were not exposed when envelope, and how does the envelope acquire its integral
the proteins were in their native conformation. Impor- spike proteins? Membrane vesicles containing the spike
tantly, the now-exposed fusion peptides are able to insert proteins pinch off from the ER and then traffic along the
into the endosomal membrane. The spike proteins then normal vesicle-mediated secretory pathway, fusing first
to the Golgi body. Depending on the particular target-
undergo further conformational rearrangements, which
ing signal on the spike protein, it may remain in the
stress the viral envelope and endosomal membrane, while
Golgi membrane, or it may undergo vesicle-mediated
also bringing them into apposition, thereby promoting
transport to the plasma membrane. By processes that are
fusion between them. This process releases the viral core
described in subsequent chapters, the viral core structure
into the cytosol. may assemble at the spike protein-containing membrane,
The low pH of the endosomal compartment is able to thereby becoming enclosed within the envelope. If this
trigger the conformational rearrangements of the spike occurs at the plasma membrane, the virus is released in
proteins because the original native conformations of the process. If it occurs at the Golgi body, the virus buds
these proteins are metastable. The concept that a virus into the Golgi lumen. It then may be released by exocytosis.
component, or indeed the entire capsid, may initially be
metastable or in a metastable arrangement is extremely
important apropos the seeming paradox that viruses as-
semble into quite stable structures at the end of one repli-
cation cycle but readily disassemble to initiate the next that of enveloped viruses. The cellular factors that trigger
replicative cycle. Do you see how the notion of initially the necessary conformational changes in these viruses are
metastable structures might explain this paradox? This is- slowly being identified, but neither the molecular nature
sue, as well as the actual conformational rearrangements of the intermediate viral structures that actually breach
of virus spike proteins and how they facilitate membrane the membrane nor the mechanisms by which they do so
fusion, is discussed in detail in the chapter on influenza are well known.
viruses (Chapter 12). Viruses that breach an endosomal membrane have
Other enveloped viruses carry out fusion at the plasma adapted to undergo their conformational transitions at a
membrane, at the time of initial attachment. In those in- pH above that which exists in lysosomes. The reason is as
stances, the fusion-promoting conformational rearrange- follows. Many cellular receptors and their ligands are des-
ments of the spike proteins are triggered by the interac- tined to be degraded in lysosomes. Toward that end, endo-
tion of the spike proteins with their receptors at the time cytic carrier vesicles transport those receptors and ligands
of attachment. from early to late endosomes, which fuse with, or mature
The mechanisms by which nonenveloped viruses breach into, lysosomes. These organelles are more acidic than
cellular membranes must be fundamentally different from endosomes, and they also contain powerful degradative
Biosynthesis and Classification of Viruses 45

enzymes that have low-pH optima. Consequently, lyso- their modes of transmission. According to this approach,
somes would be a lethal environment for many viruses. some viruses were classified as respiratory viruses, oth-
For this reason, many viruses are adapted to undergo the ers as enteric viruses, and yet others as arthropod-borne
low-pH-induced entry step under the milder conditions viruses (arboviruses). These terms are still used, but as in
that exist in the endosomal compartment, thereby avoid- the case of a pathology-based approach to classification,
ing exposure to the harsher conditions within lysosomes. they do not tell us anything about the more fundamental
The endocytic virus entry pathway confers two obvi- biological aspects of the virus.
ous advantages. (Are there others?) First, it is an efficient Many biologists might rightly argue that a system of
means for the virus to cross the plasma membrane and biological classification is useful only to the extent that it
the underlying cytoskeleton, thereby gaining access to the reflects evolutionary relatedness. But since viruses do not
deeper regions of the cell, where it might then breach a leave fossils, how might we infer evolutionary relation-
cellular membrane and replicate. Second, were an animal ships among viruses? The answer is that we do so from
virus to leave its capsid at the cell surface (like a phage morphological, from serological (i.e., immunological cross-
does), it would quickly alert the hosts immune system to reactivity), and, most critically, from molecular features of
its presence there. current viruses. Indeed, the emergence of molecular biol-
Specific examples of animal virus entry are discussed ogy made it possible to classify viruses as DNA or RNA
in detail in later chapters. For now, we can appreciate why viruses, as single stranded versus double stranded, and
the animal viruses we are about to encounter look so dif- with respect to the sizes of their genomes. Later, virus ge-
ferent from the T-even phages discussed above. nomes were sequenced, and the arrangements of the genes
within their genomes were determined. Based on this sort
of information, viruses could unambiguously be placed
THE FAMILIES OF ANIMAL VIRUSES: into distinct families, the members of which clearly are
Figure2.11 depicts a representative sampling of some, but As seen in Fig. 2.11, viruses within each of some of the
by no means all, of the virus families that infect vertebrate major currently recognized virus families not only have a
hosts. This first encounter with the animal viruses may similar size and morphology but also have genomes of
give the impression of a somewhat daunting number of common size and type (i.e., DNA versus RNA and single
individual viruses and virus families. Moreover, animal stranded versus double stranded). Most important, and
viruses seem to come in an intimidating variety of sizes not represented in the figure, is the fact that viruses within
and shapes. Thus, our purpose here (believe it or not) is to a family have a similar set of genes that are similarly ar-
simplify matters. We do this by first asserting that the ranged and expressed. Note that the viruses constituting
most fundamental problem faced by any virus is to ex- some virus families have as few as 5 genes, whereas mem-
press its genome. Then, we see that there are only a few bers of other virus families have 200 or more genes. Note
more than a handful of strategies that viruses use to solve that the arrangement of virus families in the figure is not
this most fundamental problem. Consequently, each of meant to infer evolutionary relatedness between families,
the numerous families of viruses that infect humans and since these relationships are largely unknown.
animals uses one of these strategies. But before we consider
how these points will simplify our study of the animal vi- Viral Genetic Systems: the Baltimore
ruses, we begin with a few general comments concerning Classification Scheme
the principles that have been the basis for the classification Although classifying viruses into families by the molecu-
of animal viruses in the past, as well as at present. lar criteria recounted above makes good biological sense,
In the early days of virology, when little was known we are still left with a large number of diverse virus fami-
about the biological properties of viruses, one criterion lies, which might, at first sight, appear overwhelming.
for classifying animal viruses was based on the patholo- Therefore, we ask whether there might be some rational
gies they caused. This was not a particularly useful ap- conceptual framework that might simplify our first en-
proach to virus classification, since the pathology of an counter with the large number of animal virus families
infection had little, if any, relationship with more funda- and that might also help us to keep abreast of future de-
mental biological features of viruses, in particular, the velopments in the journal literature. David Baltimore
strategies by which they express their genomes and repli- provided just such a framework, based on the notion of
cate. Also, many viruses do not cause disease. grouping virus families into classes whose members have
Another early approach to classifying viruses, which a common basic strategy for dealing with the critical issues
was favored at the time by epidemiologists, was based on of viral genome expression and replication (Baltimore,

Nucleic acid RNA

Symmetry Icosahedral Helical

Classification criteria

of capsid

Naked or Naked Enveloped Enveloped


Genome ds ds (+) ss (+) ss (+) ss (+) ss (+) ss (+) ss () ss () ss () ss () ss () ss () ss

architecture 1018 2 2 copies 3 8 2
segments segments segments segments segments

Baltimore class III III IV IV IV IV VI IV V V V V V V

Family name Reo Birna Calici Picorna Flavi Toga Retro Corona Filo Rhabdo Bunya Ortho- Para- Arena
myxo myxo
polymerase (+) (+) () () () () (+) () (+) (+) (+) (+) (+) (+)

Virion 6080 60 3540 2830 4050 6070 80130 80160 80 X 70 90120 90120 150300 50300
diameter (nm) 79014,000 85 X

Genome size 2227 7 8 7.28.4 10 12 3.59 1621 12.7 1316 13.521 13.6 1620 1014
(total in kb)

Figure2.11 Some of the major families of animal viruses, grouped according to their shared properties. The most
fundamental shared distinguishing characteristics of members of a family are the organization of their genomes and
their patterns of gene expression. Other shared characteristics are shown as well. The Baltimore classification system is
superimposed on the figure (see the text). The figure is not meant to indicate evolutionary relationships between virus
families, since these are largely unknown. Adapted from M. H. V. van Regenmortel etal. (ed.), Virus Taxonomy Classifi-
cation and Taxonomy of Viruses: Sixth Report of the International Committee on Taxonomy of Viruses (Springer-Verlag,
Vienna, Austria, 1995), as appears in S. J. Flint, L. W. Enquist, R. M. Krug, V. R. Racaniello, and A. M. Skalka, Principles of
Virology: Molecular Biology, Pathogenesis, and Control, 2nd ed. (ASM Press, Washington, DC, 2003).

1971). Baltimores original classification scheme defined RNA as a template for transcription. Moreover, reovirus
six such fundamental strategies and six corresponding vi- double-stranded RNA genomes do not have mRNA activ-
rus classes (Fig. 2.12). We might now add a seventh class, ity of their own. Yet if the virus could somehow transcribe
to include the hepadnaviruses, which do not easily fit into the gene that encodes its own RNA-dependent RNA poly-
any of the original six classes. Then, each of the recog- merase (or transcriptase), then its predicament would be
nized virus families can be grouped into one of the seven solved. The reason for this is based on a point that we de-
Baltimore classes. But see also Box 22.1. veloped at length earlier in the chapter, i.e., the fact that
To appreciate the rationale for Baltimores approach, virus mRNA is translated by the cells own translation
consider as an example the particular strategy adapted by machinery. Thus, a virus can generate any special func-
the reoviruses, which have double-stranded RNA genomes tion that it might require, provided it can transcribe the
(Fig. 2.11). We begin by noting that there is no enzymatic viral gene that might encode the needed function. In the
machinery in the cell that can replicate double-stranded case of the reoviruses, they cannot use a host RNA poly-
RNA. Thus, if the reovirus is to replicate, it will have to merase to make the necessary mRNA. So how do reovi-
produce its own replicase. However, to do this, the virus ruses escape from the circular dilemma we have posed?
must first transcribe the viral gene that might encode its The answer is that these viruses bring the necessary
replicase. But the dilemma for the virus continues because transcriptase activity into the cell, contained within the
there also is no cellular enzyme that can use double-stranded virus particle. Those virus particle-associated transcriptase
Biosynthesis and Classification of Viruses 47


Icosahedral Complex

Naked Enveloped Naked/ Enveloped

enveloped (cytoplasmic)
ss linear ss ds ds ds ds ds ds ds
(+) or () circular circular circular linear circle linear linear covalently
gapped joined ends


Parvo Circo Polyoma Papilloma Adeno Hepadna Herpes Irido Pox

() () () () () (+) () () (+)

1826 1226 40 55 7090 42 150200 125300 170200

X 300450

5 1.82.3 5 78 3638 3.2 120200 150350 130280

Figure2.11 Continued

molecules were produced during the preceding cycle of Figure2.12 The Baltimore classification scheme, depicting the
infection and were incorporated into the progeny reovirus relationship of viral genomes to the mRNAs they produce. Viruses
are classified based on the nature and polarity of their genomes. Each
particles during their maturation. We can now appreciate class uses a different strategy to generate its mRNA. This provides a
what is meant by patterns of strategies of virus gene ex- fundamental distinction between classes, since all virus mRNAs are
pression and how these might vary from one virus family translated by the cellular translation machinery. An example of each
to another. Reoviruses are the only known double- class is indicated. From D. Baltimore, 1971, as adapted in S. E. Luria, J.
J. Darnell, D. Baltimore, and A. Campbell, General Virology, 3rd ed.
stranded RNA virus family that infects vertebrates. Bacte- (John Wiley & Sons, New York, NY, 1978), with permission.
riophage j6, a double-stranded RNA phage that naturally
infects Pseudomonas phaseolicola, is discussed in Box 5.8.
Double-stranded RNA viruses are class III viruses in the Class II
Baltimore classification scheme (Fig. 2.12). Parvovirus
Class VI
The class IV viruses in the Baltimore scheme provide RNA DNA DNA
an interesting contrast to the reoviruses, as well as to the
Class I
class V viruses. Class IV and class V viruses each have Vaccinia virus
single-stranded RNA genomes, but with one fundamen-
tal difference. To appreciate that key difference, first note RNA RNA mRNA Class III RNA
that at any site along a double-stranded DNA molecule, Reovirus
only one of the DNA strands usually serves as a template Class IV
Poliovirus Class V
for transcription. That strand, which is complementary in Vesicular stomatitis virus
base sequence to the mRNA, is referred to as the minus
strand, since by convention, mRNA is always of the plus RNA

Box 2.9 common colds), hepatitis A virus, rubella virus, West Nile
virus, foot-and-mouth disease virus, yellow fever virus,
dengue virus, and the severe acute respiratory syndrome
When single-stranded RNA viruses generate progeny ge-
(SARS) coronavirus. Class V viruses include other favor-
nomes, as well as mRNAs, there is no true double-stranded
RNA intermediate per se. Instead, each single-stranded
ites such as influenza virus, measles virus, mumps virus,
RNA template simultaneously transcribes several stag- rabies virus, and Ebola virus.
gered nascent single-stranded complementary RNA Class I viruses are the double-stranded DNA viruses. In
chains. These replicative intermediates do not collapse principle, transcription of their genomes should not pres-
into double-stranded RNA forms. Perhaps single-stranded ent any special problem, since the enzymatic machinery
RNA viruses evolved this mode of replication to avoid necessary for the transcription and replication of double-
inducing interferon (see Chapter 4). How then might stranded DNA already exists in the cell. Since this enzy-
the reoviruses avoid inducing interferon when replicat- matic machinery is located exclusively in the nucleus,
ing their double-stranded RNA genomes? The answer is most class I viruses indeed replicate in the nucleus. The
quite astonishing, at least to me (see Chapter 5). poxviruses (Chapter 19) are the only well-studied excep-
tion to this generalization.
Although most class I viruses use cellular enzymes to
mediate their transcription and replication, they never-
theless have a special concern, which stems from the fact
sense. The other DNA strand has the same base sequence that crucial host enzymes that mediate DNA metabolism
as mRNA. Hence, like the mRNA, it is of the plus sense. are efficiently expressed in cells only when they are in the
Returning now to the class IV and class V viruses, class IV S phase of the cell cycle, and most cells in the vertebrate
viruses have single-stranded RNA genomes that are of the host are not cycling. To solve this dilemma, many class I
plus sense. That is, their genomes are of the same sense as viruses have evolved the ability to induce nondividing or
the mRNA that they generate. In contrast, class V viruses resting (a misnomer) cells to enter into S phase. Indeed,
have genomes that are of the minus sense. That is, their the ability of these viruses to induce resting cells to enter
genomes are complementary in base sequence to the into S phase underlies the ability of some of them to in-
mRNA that they transcribe. duce tumors in the laboratory in some instances and in
Both class IV and class V viruses must generate com- their human and animal hosts, as well, in other instances.
plementary RNAs in order to produce mRNAs and prog- Class I viruses include papillomaviruses (responsible for
eny viral genomes (Box 2.9). And as in the case of the cervical carcinoma and common warts), adenoviruses,
class III reoviruses, there is no enzymatic machinery made herpesviruses (including the herpesviruses responsible
by the cell that can either transcribe or replicate RNA. for genital herpes, chicken pox and shingles, and mono-
How then do class IV and class V viruses manage to carry nucleosis), and smallpox virus.
out their transcription and replication? Class II viruses have single-stranded DNA genomes.
For class IV viruses the solution is relatively straight- The parvoviruses, which include some of the smallest
forward. Since their single-stranded RNA genomes are of known viruses, are the only single-stranded DNA viruses
the plus sense, their genomes may be directly translated known to infect animals. Interestingly, a subgroup of the
on host cell ribosomes to yield the required virus-encoded parvoviruses, the dependoviruses (also known as adeno-
transcriptase and replicase activities that they encode. associated viruses), are able to replicate only in cells that
Indeed, the first step in the replication cycle of class IV are coinfected with an adenovirus or a herpesvirus, which
viruses, after uncoating, is the translation of their genomes serves as a helper for the dependoviruses. That is, wild-
on cellular ribosomes. type dependoviruses require one of several different unre-
What then of the class V viruses? Since their single- lated helper viruses to promote their replication. Depen-
stranded RNA genomes are of the minus sense, their ge- doviruses may encapsidate either a plus-DNA strand or
nomes cannot be translated. You may have correctly rec- a minus-DNA strand. In contrast, autonomous parvovi-
ognized that their situation is essentially the same as that ruses, which do not require a helper virus, usually package
of the class III reoviruses. Not surprisingly, then, their so- only minus-DNA strands. Regardless, transcription of
lution is the same as that of the reoviruses. That is, the parvovirus DNA by the host cell RNA polymerase requires
required polymerase activity enters the cell as a compo- that the parvovirus first generate a double-stranded DNA
nent of the class V virus particle. replicative form; this step is catalyzed by host cell enzymes.
Class IV viruses include such well-known examples as And, as you might then have supposed, class II viruses,
poliovirus, the rhinoviruses (responsible for many of our like class I viruses, replicate in the nucleus.
Biosynthesis and Classification of Viruses 49

Class VI viruses are the remarkable retroviruses. These the hepadnaviruses should constitute a separate Baltimore
viruses have single-stranded RNA genomes of the plus class of their own, see Box 22.1.
sense. However, unlike class IV plus-sense RNA virus ge- Before moving on, note that the Baltimore classifica-
nomes, retrovirus plus-sense RNA genomes do not serve tion scheme is not a formal means for classifying viruses.
as mRNAs. Instead, in a process catalyzed by a virus- Instead, it provides a rational basis for grouping virus
encoded enzyme (the reverse transcriptase), the retrovi- families, based on their strategies for gene expression.
rus RNA genome is reverse transcribed to generate Knowing where any virus fits in the Baltimore scheme en-
double-stranded DNA, which then is integrated into a ables one to approach the journal literature discussing
cellular chromosome. The integrated viral DNA, referred that virus with a wealth of background information about
to as a provirus, is transcribed by the cellular pol II RNA the replication strategy of that virus already at hand.
polymerase to generate all retroviral mRNA molecules From the experience of many years of teaching virol-
and all retroviral progeny RNA genomes as well. As you ogy, I am certain that I have not overestimated the value
may already have surmised, the retroviral reverse tran- of the Baltimore scheme as a didactic approach to the
scriptase enters the cell with the viral RNA genome. There animal viruses. On the other hand, be aware that there are
is much to say about this extraordinary virus family that important features other than the general patterns of
will have to wait (Chapters 20 and 21). For the moment, transcription that distinguish one virus family from an-
note that this family includes the RNA tumor viruses (in- other within a Baltimore class, as well as features that dis-
cluding the Rous sarcoma virus), studies of which led to tinguish between members of the same family. Apropos
enormous insights into cellular growth control and can- that point, think of the vertebrate host and its cells and
cer, and HIV, the causative agent of AIDS. tissues as the viruss environment. Also, think of the highly
The reproductive cycle of the hepadnaviruses, which effective innate and adaptive immune responses of the
include hepatitis B virus and its relatives, was determined host (Chapter 4), as well as the defensive measures that
after the Baltimore scheme was first put forward. Although individual cells are endowed with, as factors in that envi-
hepadnaviruses have double-stranded DNA genomes, ronment. While these host factors would seem to make
their replication cycle differs fundamentally from that of the vertebrate host a not particularly hospitable environ-
the class I viruses, and consequently, this virus family ment for any would-be microbial invader, once immersed
might now be seen to merit a Baltimore class of its own within their hosts, viruses are exquisitely endowed to
(i.e., as class VII viruses). In brief, the hepadnavirus repli- thrive. Like any other successful organism, viruses are
cation cycle is as follows. Cellular RNA polymerase tran- adapted to their environments. Virus adaptations enable
scribes the viral circular double-stranded DNA genome these parasites to enter permissive cells within their hosts,
into a longer-than-genome-length RNA molecule. This release and express their genomes, produce and release
RNA then serves as mRNA that is translated into capsid progeny viruses, and transmit the infection to new hosts
proteins and, also, an enzyme that has reverse transcriptase before they might be cleared by host immune functions.
activity. The next steps are truly remarkable. The long The variety of adaptations that different viruses have
RNA, in association with the reverse transcriptase, is then evolved, both within and between families, to carry out
packaged into capsids. Within the capsids, the RNA is re- each of the steps of their unique lifestyles, as well as to
verse transcribed to generate progeny double-stranded evade and subvert host defenses, is one of the most fasci-
DNA genomes. The template RNA is degraded in the pro- nating aspects of virology.
cess. Note that hepatitis B virus is a major human patho-
gen. It is a major cause of hepatitis, and it is responsible Suggested Readings
for over a million cases of fatal hepatocellular carcinoma Baltimore, D. 1971. Expression of animal virus genomes. Bacte-
(a liver cancer) per annum, primarily in Asia. riol. Rev. 35:235241.
Since hepatitis B virus and the retroviruses use reverse Cairns, J., G. S. Stent, and J. D. Watson (ed.). 2007. Phage and the
transcription in their replication, we recap the essential Origins of Molecular Biology: the Centennial Edition. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY. (This is the
differences in their replication strategies. Hepatitis B virus latest edition of the original 1966 publication.)
uses an RNA intermediate to replicate its DNA genome, Ptashne, M. 2004. A Genetic Switch: Phage Lambda Revisited,
whereas retroviruses use a DNA intermediate to replicate 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring
their RNA genomes. For further discussion of whether Harbor, NY.
Hematogenous and Neural Dissemination
The Placenta and the Fetus

Modes of Virus Infection

Acute Infections
Persistent Infections
Slow Infections and Disease
Chronic Infections
Latent Infections
Transmissible Spongiform Encephalopathies:

Our purpose in this chapter is to provide an overview of the various ways
different viruses may interact with their hosts invivo. We begin by consid-
ering how different viruses are transmitted within their host populations,
after which we discuss the means by which they then disseminate within
an infected host. Next, we consider the diverse patterns of infection within
the host, paying particular attention to the distinctions between rapidly
resolving acute infections and the several kinds of persistent infections, in
which the virus may persist for the lifetime of the host.
But again, this chapter is an overview, in which several key points are
illustrated by a relatively few selected examples. Thus, some medically im-
portant viruses, such as the human immunodeficiency virus (HIV), the
influenza viruses, and Epstein-Barr virus (EBV), may seem to be men-
tioned all too briefly here, and others, such as some notable hemorrhagic
fever viruses, are not mentioned at all. However, all of these viruses are
considered at greater lengths in later chapters.
One point to be made before moving on is that Chapter 4, which fol-
lows, is meant to provide an immunology background for all immunity-
related issues that might come up anywhere in the text. Thus, the informa-
tion contained in Chapter 4 would certainly make the present chapter
more accessible for many readers. Nevertheless, I chose to place Chapter 4
after the present chapter for two noncritical reasons. First, I wanted to get
the reader further into the subject of virology before taking a big detour
into immunology. Second, the information in the present chapter does
provide an important context for the material in the following chapter. So,
which chapter to place first was actually a toss-up. Regardless, the real point
is this: please read Chapter 4 first if that works better for you, or perhaps you
might skim or read Chapter 4 along with the present chapter.

Viruses may cause disease only after they have gained access to susceptible
cells and tissues of the host. But before disease symptoms might emerge,

Modes of Virus Infection and Disease 51

the virus must replicate at the initial or primary site of all too familiar symptoms. Other viruses, such as measles
infection, and in the case of some viruses, the infection virus and varicella-zoster virus (VZV) (a herpesvirus re-
must disseminate to secondary sites as well before disease sponsible for chicken pox and shingles), also enter via the
might emerge. respiratory tract but then disseminate to secondary sites
Most viral infections are transmitted via the respiratory of infection.
route (i.e., via inhalation). The reasons why many viruses The skin is an inherently more formidable barrier to
exploit the respiratory tract as a site of entry become clear viruses than the respiratory tract since (i) it is several cell
when we consider the following points regarding the re- layers thick; (ii) the outermost layer is comprised of tough,
spiratory tract. The purpose of the respiratory tract is to dead, highly keratinized cells; and (iii) the skin surface is
exchange oxygen for CO2. To accomplish this task, hu- acidic due to the fatty acids secreted by the sebaceous
mans respire about 6 liters of air per min. Moreover, the glands (Fig. 3.1). Still, some viruses do invade the skin by
surface area of the respiratory mucosa (mucosa refers to taking advantage of breaks caused by injury. When the in-
the epithelial layer lining the respiratory, digestive, and tegrity of the skin is compromised, these viruses might
urogenital tracts) must be large. Indeed, it is about 140m2. target the mitotically active keratinocytes of the germinal
For comparison, the entire surface area of the skin is a stratum or other cells of the basal layer, as well as cells
mere 2m2. Furthermore, the respiratory mucosa is deli- beneath the skin. (Keratinocytes are the epidermal cells
cate and richly vascularized. that synthesize keratin and undergo characteristic changes
Since the respiratory tract has features that make it a as they move upward from the basal layers to the cornified
singularly attractive portal of entry for pathogens, Nature layers of the skin.) Viruses that replicate in the germinal
has seen fit to endow it with an impressive array of defen- stratum include the herpes simplex viruses (HSVs) (the
sive mechanisms (Chapter 4). Nevertheless, some viruses cause of cold sores, fever blisters, and genital herpes) and
have evolved effective countermeasures that enable them some poxviruses. The papillomaviruses (the cause of com-
to yet exploit the respiratory tract as an entranceway into mon warts and cervical carcinoma) have a more complex
the host. replication cycle, in which successive stages of their repli-
The human rhinoviruses (the major cause of the com- cation take place in different layers of the epithelium or
mon cold) and influenza viruses are among the viruses the genital mucosa.
that exploit the respiratory tract as their portal of entry. In Other viruses initiate infection across the skin by mak-
the case of these two virus groups, all of their replication ing use of insect vectors. Yellow fever virus provides a
may occur in the respiratory mucosa, accounting for their prominent example, as noted in Chapter 1. Other notable

Figure3.1 A schematic representation of the skin. Note the multilevel nature of the epidermis, which contains the
stratum corneum, consisting of several rows of dead keratinized cells. The stratum corneum in turn overlies the stra-
tum lucidum, the stratum granulosum, and the basal layer (stratum germinativum). Also note that the epidermis is
purely a cellular structure, without blood vessels or nerves. Below the epidermis are the basement membrane and the
dermis. The latter contains blood vessels, lymphatic vessels, macrophages, and fibroblasts (which form the fibrous
tissue and the matrix or ground substance). Adapted from F. Fenner et al., The Biology of Animal Viruses, Academic
Press, New York, NY, 1974, with permission.

Sebaceous gland
Stratum corneum
Stratum lucidum
Stratum malpighii Stratum granulosum Basement
Stratum germinativum membrane

Lymphatic vessel

Subcutaneous fat
Hair follicle
Blood vessel

Box 3.1 mediators, including phagocytes and complement, and

also enzymes (e.g., amylases) that may destroy viruses and
virus-infected cells. But then, as you might have expected,
In Chapter 2, we discussed criteria that were used for
viruses that exploit the oropharynx as a site of entry have
classifying viruses before the molecular biology era,
noting that one early classification scheme was based
evolved countermeasures to subvert these host defenses.
on mode of transmission. This approach to virus clas- Examples of these viruses include HSV (oral herpes), EBV
sification gave rise to the designations respiratory virus, (a herpesvirus that causes infectious mononucleosis), and
enteric virus, and arbovirus (derived from arthropod- some papillomaviruses.
borne virus). The term arbovirus is still sometimes The oral cavity leads to the gastrointestinal tract, which
used today to refer to a vertebrate-infecting virus that is includes the stomach, small intestine, large intestine, rec-
transmitted by an arthropod vector. But on first encoun- tum, and anal canal. The gastrointestinal tract too is a
tering this term, one might well conclude that it refers to harsh environment for most viruses. Wild fluctuations in
a virus that actually spends its life cycle in an arthropod pH are found there, as the stomach is acidic (epithelial
host. Contrary to that expectation, arboviruses are actu- cells lining the gastric glands secrete hydrochloric acid, re-
ally dependent on their vertebrate hosts, as follows. The ducing the pH of the stomach to as low as 2), whereas the
arthropod becomes infected upon taking a blood meal intestine is alkaline. In addition, there are digestive en-
from an infected vertebrate. The virus then replicates in zymes (e.g., proteases, amylases, and lipases) and bile salts
the arthropods gut epithelium, from which it dissemi- (which emulsify fats and destroy viral envelopes), and the
nates to the salivary gland, where it mixes with the saliva. luminal surfaces of the tract are covered with mucus that
Thus, a bite can transmit the virus to the next vertebrate also harbors antibodies and phagocytes.
victim. Arboviruses generally produce only persistent The mucosa of the small intestine also includes the
infections in the arthropod host, while producing either Peyers patches, which are focal aggregates of organized
acute or persistent infections in the vertebrate host. There lymphoid tissue (Fig. 3.2 and 3.3). Like lymph nodes, Pey-
are genuine insect viruses, such as the baculoviruses (not ers patches are sites at which antigens are presented to the
discussed further in this text), which indeed spend their
effector cells (i.e., B and T lymphocytes) of the adaptive
entire life cycles in the arthropod host.
immune system, thereby activating an adaptive immune
response against the invader (Chapter 4). Lymphocytes
enter Peyers patches, as well as lymph nodes, from the
blood circulation. However, whereas antigen enters lymph
examples include West Nile virus, Japanese encephalitis nodes from the lymphatic vessels that drain the infected
virus, dengue virus, and St. Louis encephalitis virus, sites, there are no lymphatics that drain into Peyers
among others (Box3.1). patches. How then does antigen enter Peyers patches? The
Finally, we should acknowledge that the skin may be answer is through specialized cells of the Peyers patches
penetrated and infection may occur as a direct result of called M cells, which specialize in transporting antigens
human activity. This may happen inadvertently during from the intestinal lumen to the underlying cells of Peyers
medical procedures, such as blood transfusions. Such patches. In order to carry out their task, M cells have a
medically induced events are said to be iatrogenic (from very thin cytoplasm, and they transport antigens by tran-
iatros, the Greek word for physician). More commonly, scytosis, delivering them virtually intact to the underlying
penetration of the skin occurs as a consequence of the lymphoid tissue. (Transcytosis is a process that occurs in
sharing of hypodermic syringes by intravenous drug us- polarized epithelial and endothelial cells [see below]. Li-
ers. Note that this is a major route of transmission for gands are taken up in endocytic vesicles on one surface of
HIV and for hepatitis B and C viruses (HBV and HCV). the polarized epithelium. Then, the cargo is transported
The mucosa of the alimentary tract, which lines the to the opposite surface of the epithelium by a vesicle-
entire cavity running from the mouth to the anus, is an- mediated process. At the opposite surface, the cargo is re-
other major portal of virus entry. As you might expect, the leased intact by fusion of the transport vesicle with the
food we eat and the fluids we drink may be contaminated plasma membrane. As noted above, M cells are highly spe-
with a variety of pathogens. Accordingly, the alimentary cialized for transcytosis.)
tract, like the respiratory tract, is armed with a formidable Unfortunately for the host, some pathogens subvert
array of defensive measures. These begin in the oral cavity, the very properties of M cells that enable those cells to
where the alimentary tract is protected by an epithelium transport antigen to the underlying Peyers patch, doing
that is several cell layers thick. Moreover, the oral cavity is so to cross the intestinal epithelium. For example, when
bathed in saliva, which contains several important immune reoviruses are taken up by M cells and transferred by
Modes of Virus Infection and Disease 53

Microvillus Apical side

Left subclavian vein
Right subclavian vein Thymus
Lymph node Heart
Enterocyte Thoracic duct
Tight junction
Kidney Spleen
M cell pocket
Peyers patch in
small intestine
Large intestine
Bone marrow

M cell

membrane Figure3.3 Distribution of lymphoid tissues in the body. Note the
concentration of lymph nodes in the region of the nasopharynx, the
Figure3.2 Drawing of a section of the intestinal epithelium containing lungs, and the gastrointestinal and urogenital tracts. Modified from
an M cell connected to two enterocytes (i.e., simple columnar epithe- R. Coico, G. Sunshine, and E. Benjamani, Immunology, A Short Course,
lial cells found in the small intestines and colon). Lymphocytes and 5th ed. (John Wiley & Sons, Hoboken, NJ, 2003), with permission.
macrophages move in and out of the invagination of the M cell at its
basolateral side. Notice the thinness of the M-cell cytoplasm. Although
M cells are present for defensive purposes, a variety of microbial
pathogens exploit M cells to cross the intestinal epithelium. Adapted Note that the rhinoviruses, which cause common colds,
from A. Siebers and B. B. Finlay, Trends Microbiol. 4:2228, 1996, and are picornaviruses. As such, they are closely related to the
B. Alberts et al., Molecular Biology of the Cell, Garland Publishing, other picornaviruses enumerated above. Yet the rhinovi-
New York, NY, 1994, with permission.
ruses are sensitive to low pH, and they have adapted to
infecting the respiratory tract rather than the alimentary
transcytosis to underlying Peyers patches, these viruses then tract (Chapter 6). Thus, closely related viruses may use
use the occasion to disseminate through the host via the different portals of entry and different target cells in which
circulation. In contrast, the related rotaviruses replicate in, to replicate.
and destroy, M cells. This brings about the inflammation of HIV, when transmitted via anal intercourse, can infect
the gut and diarrhea characteristic of rotavirus infections. the lower gastrointestinal tract without having to traverse
Despite the above two examples of viral invasion via the upper respiratory tract and all the inherent hazards it
M cells, the epithelial cells and white blood cells of the poses to a virus. Under these circumstances, HIV can in-
gastrointestinal mucosa are the principal cells that viruses fect dendritic cells in the epithelium of the anus, another
target in that tissue. The corresponding cells of the instance in which a virus subverts an immune effector cell
oropharynx likewise are the principal cells targeted by for its own purposes. Dendritic cells are called profes-
viruses at that site. sional antigen-presenting cells (Chapter 4). They con-
Other virus types that initiate infection in the gastroin- tinually sample the environment at sites of pathogen entry.
testinal tract include adenoviruses, parvoviruses, picorna- Upon being activated by a pathogen, an immature senti-
viruses (a family that includes poliovirus, hepatitis A vi- nel dendritic cell matures into an effector dendritic cell
rus, and enteroviruses), coronaviruses, and the Norwalk and migrates to a draining lymph node, where it presents
calicivirus (renamed norovirus). It is not entirely clear antigen to a nave T cell (i.e., a T cell that has not previ-
how these viruses might breach the defensive barriers of ously encountered its specific antigen and been activated),
the gastrointestinal tract. Yet each of these viruses is resis- thereby activating an adaptive immune response (Chap-
tant to low pH, to gastrointestinal proteases, and to the ter 4). In the case of HIV infection, when the dendritic cell
emulsifying action of bile detergents. (With the exception delivers HIV to the lymph node, in point of fact it is deliv-
of the coronaviruses, none of these viruses are enveloped.) ering the virus to the very site at which its principal host

cell, the CD4 T lymphocyte, is concentrated. Importantly, of the blinking response. In addition, tears contain IgA
the tropism of HIV for CD4 T cells is the major determi- antibodies (Chapter 4). Moreover, the conjunctiva and
nant of the pathogenesis of acquired immunodeficiency sclera (i.e., the fibrocollagenous covering of the eye) are
syndrome (AIDS) (Chapter 21). Note that the virus may protected by associated lymphoid tissue. Nevertheless,
not need to be internalized by the dendritic cell during several viruses (adenoviruses, herpesviruses, and entero-
these events (a point discussed at greater length in viruses in particular) are able to initiate infection via the
Chapter 21). A similar course of events may occur during conjunctival route. In these instances, the opportunity for
vaginal intercourse (see below), but the risks of infection infection is increased if the mucosa is injured by abrasion.
are greater during anal intercourse because of the greater The source of the infection is usually contact of the
risk of tissue damage there. conjunctival tissue with contaminated fingers, objects, or
The urogenital tract can be invaded by several viruses, pool water. Conjunctival infections generally remain lo-
despite the fact that it too manifests several barriers to in- calized in the conjunctiva, resulting in an inflammatory
fection. Regular urination is one such barrier, since it flushes condition referred to as conjunctivitis, or more colloqui-
out most pathogens before they might initiate infection. ally as pink eye. However, herpesvirus infections of the
In addition, urine contains immunoglobulin A (IgA) an- cornea usually spread to sensory neurons that innervate
tibodies (the type of antibodies secreted by mucosal lym- the cornea. These viruses then establish lifelong persistent
phoid tissue [Chapter 4]), and it is also slightly acidic. infections in these neurons, which rarely are serious (see
Sexual intercourse is the principal means by which sev- below). Enterovirus 70infections of the eye seldom dis-
eral medically important viruses invade the urogenital seminate, but when they do, they may lead to a potentially
tract. The tears and abrasions that result from normal serious paralytic condition.
sexual activity may facilitate infection via this route. We
noted above that HIV can be transmitted via anal sex.
However, the mucosal surfaces of the female genital tract ROUTES OF DISSEMINATION
are actually the primary portal by which HIV is spread. Hematogenous and Neural Dissemination
That is to say, heterosexual transmission is the major route As we have seen, viruses invade the body through breaks
by which HIV is spread worldwide (Chapter 21). Other in the skin or by infecting the mucoepithelial membranes
viruses that invade the urogenital tract include HSVs, that line the respiratory tract, oropharynx, gastrointesti-
papillomaviruses (the cause of genital warts and cervical nal tract, urogenital tract, and eyes. In many instances, the
cancer, as noted before), and HBV. virus remains at the initial site of infection, and any symp-
For HIV to be transmitted via the female urogenital toms that may arise emanate from that site. For example,
tract, the virus must cross the genital mucosa so that it the rhinoviruses, which cause common colds, invade the
may make its way to a lymph node, where it finds its prin- upper respiratory tract and do not disseminate from that
cipal target cell, the CD4 T lymphocyte. The mechanism initial site of infection. Other viruses may replicate at the
by which HIV accomplishes these steps is not well under- primary site but then disseminate to secondary sites. In
stood. One possibility is that HIV crosses the mucosa via these cases, symptoms might result from tissue damage at
lesions caused by other infections. However, experiments the secondary site.
with rhesus macaques show that mucosal lesions are not The lymphatic system and the blood circulation are the
necessary for infection with the related simian immuno- major conduits for virus dissemination in the body. As vi-
deficiency virus, which appears to be able to cross an in- rus particles are released from infected cells into the extra-
tact epithelium. HIV might cross the urogenital epithe- cellular space, they may then enter a local afferent lym-
lium by M-cell-mediated transcytosis. The virus might phatic vessel, which drains the infected site. These afferent
then exploit the fact that Langerhans cells, which are the lymphatics normally deliver antigen and dendritic cells to
dendritic cells at that site, lie directly under the mucosa. the nearest draining lymph node (Fig. 3.3). Although free
That is, HIV might use the Langerhans cells for transport virus particles may drain to the lymph nodes, some vi-
to the lymph node (see above). Coincidentally, these ruses may be phagocytosed by sentinel dendritic cells at
Langerhans cells might be especially well suited to this the infected site (Chapter 4) and then subvert those phago-
task because they express a particular lectin, DC-SIGN, cytes to transport them to the nearest lymph node. In that
that binds the HIV envelope glycoprotein gp120. As noted regard, recall that during vaginal and anal intercourse,
above, the virus may not need to be internalized by the HIV interacts with dendritic cells in the epithelium. Those
dendritic cell during these events. dendritic cells might then deliver the virus to lymph
The cornea and conjunctiva (i.e., the inner surface of nodes, where the virus is able to infect its principal target
the eyelid) are protected by the continual cleansing action cell, a CD4 T cell.
Modes of Virus Infection and Disease 55

Other viruses, such as several enteric viruses of the generally inaccessible sites, such as the central nervous
picornavirus and reovirus families, bind to receptors on system and brain. When many organs become infected,
M cells, which then transport them to underlying Peyers the infection is said to be systemic.
patches of the lymphatic system. But by whatever means Virus in the blood may be free, or it may be associated
virus particles enter into the lymphatic circulation, they with lymphocytes or macrophages (macrophages are cells
are delivered to local draining lymphoid organs via affer- that are specialized to phagocytose microbial pathogens
ent lymphatic vesicles. To appreciate what happens next, [Chapter 4]). Most viruses that are taken up by mac-
we have to briefly consider the lymphatic circulation. rophages are likely to be destroyed by them. However,
Blood-borne lymphocytes enter the lymph nodes through some viruses may replicate in macrophages, using them as
postcapillary venules. They return to the blood circula- a metaphorical Trojan horse to mediate their transport to
tion by leaving the lymph nodes via efferent lymphatic other tissues of the body. Indeed, some viruses may enter
vesicles, which converge in the thoracic duct (Fig. 3.3). the central nervous system by infecting migrating mac-
The thoracic duct empties into the vena cava, the vessel rophages, while other viruses may use lymphocytes for
that returns blood to the heart, and thus returns lympho- that purpose.
cytes back to the blood circulation. This lymphatic circu- While some viruses can invade the central nervous sys-
lation is crucial in immunity, because it enables antigen, tem or brain via the blood circulation, other viruses can
antigen-presenting cells, and lymphocytes to come to- spread from the primary site of infection to the central
gether in lymphoid organs to initiate an adaptive immune nervous system by infecting the endings of neurons that
response (Chapter 4). But some viruses exploit the lym- innervate or are near the infected site (Fig. 3.4). Indeed,
phatic circulation, such that the lymphatic system is the neural spread is second only to the blood circulation as
major route by which viruses invade the bloodstream. the means by which viruses disseminate in the host.
There are still other means by which a virus may enter In order for a virus to replicate in a neuron, it must
the blood circulation. For instance, infection often leads make its way to the cell body of the neuron, since protein
to a local inflammatory response, the extent of which de- synthesis does not occur in the extended processes of these
pends on several factors (Chapter 4). Inflammation is cells. Viruses access the neuronal cell body and spread along
characterized by the dilation of local blood vessels, ren- neurons by microtubule-associated fast axonal transport.
dering them more permeable to the cellular and humoral Indeed, neural spread is a defining characteristic of infec-
mediators of immunity. Although the inflammatory re- tions by certain viruses, including rabies virus, HSV, and
sponse is a host defensive measure, the increased perme- VZV. For poliovirus, neural invasion occurs inadvertently
ability of capillaries at the infected site may enable some and rarely, but it is a critical factor in the pathogenesis of
viruses to enter the blood circulation. Viruses also may paralytic poliomyelitis (Chapter 6).
enter the blood by replicating in the endothelial cells that Only olfactory neurons have nerve endings that are ex-
line the blood capillaries. Arboviruses may be transmitted posed to the environment. Accordingly, neurotropic vi-
to the blood by the bite of an arthropod vector. Also, virus ruses generally must replicate in nonneuronal cells before
entry into the blood circulation may be iatrogenic or, they are able to access neurons. Since the epithelial muco-
more commonly, may occur as a result of needle sharing sas are usually the primary sites of virus infection, the first
by drug users, as in the cases of HIV and hepatitis B and C neurons to be infected are commonly components of the
viruses (see above). peripheral nervous system. Rabies virus, which is trans-
The initial appearance of virus in the blood is referred mitted by an animal bite, is generally introduced into
to as a primary viremia. If the virus should then replicate muscle tissue that is rich in nerve endings. Thus, it can then
in a secondary site of infection, such as the endothelial invade and ascend peripheral neurons, eventually cross-
lining of blood vessels or the liver, spleen, or lungs, then ing their synapses by transsynaptic spread. Virus spread
progeny virus from the secondary site also may enter the can continue in this way, with the virus eventually making
blood circulation, resulting in a secondary viremia. Sec- its way to the central nervous system or brain.
ondary sites of infection, such as the liver, may support HSV and VZV (the neurotropic herpesviruses) typi-
extensive virus replication. Accordingly, secondary vire- cally become latent in the cell bodies of peripheral neu-
mias usually lead to much greater levels of virus in the rons (see below and Chapter 18). When reactivated, they
blood than primary viremias. Measles virus provides a usually make their way back down the neuron and rein-
good example of a virus that gives rise to acute primary fect the original epithelial site of infection. Only rarely
and secondary viremias (see below and Chapter 11). The will these herpesviruses also move in the retrograde di-
high levels of virus in the blood resulting from the sec- rection to the central nervous system. Thus, the more
ondary viremia may enable the virus to infect yet other common direction of herpesvirus movement following

Dorsal root ganglion Spinal cord

Sensory pseudounipolar neuron

Motor neuron

end plate

Schwann cell


og r a
Perineural lymphatics de

Endoneural space

Figure3.4 Some viruses can spread from the primary site of infection to the central nervous system by infecting the
endings of neurons that innervate or are near the infected site. Directed transport within the neuron is defined as an-
terograde (movement from the minus to the plus ends of microtubules) or retrograde (movement from the plus to the
minus ends of microtubules). Adapted from R. T. Johnson, Viral Infections of the Nervous System, Raven Press,
New York, NY, 1982, with permission.

reactivation results in a cold sore, whereas the rarer retro- Infection of the brain may give rise to other distinct
grade transport leads to lethal viral encephalitis (defined and important clinical conditions, in particular aseptic
below). (Note that anterograde versus retrograde neu- meningitis and myelitis. Beginning with aseptic meningi-
ronal transport refers to the direction of microtubule tis, many viruses, including some enteroviruses (Chapter 6)
movement. Anterograde is from the minus ends to the plus and HSV type 2 (HSV-2) (sometimes referred to as genital
ends of microtubules. Thus, movement toward the cell herpes [Chapter 18]), infect the meninges (the lining of
body is called retrograde transport and movement toward the brain and the brain stem), where they give rise to viral
the synapse is called anterograde transport, irrespective of or aseptic meningitis. (Aseptic meningitis is a misnomer
the direction with regard to the central nervous system.) used to distinguish viral meningitis and meningitis from
Encephalitis refers to infection of the actual neurons of noninfectious causes [e.g., a side effect of medications] from
the brain and brain stem, as distinct from their supportive meningitis caused by other infectious agents.)
tissue. Rabies virus provides a well-known example of a When viruses infect the neurons of the spinal cord,
virus that gives rise to encephalitis in the brain. Note that they can give rise to a condition known as myelitis. For
rabies encephalitis is almost invariably fatal, whereas other example, poliovirus, an enterovirus of the picornavirus
instances of viral encephalitis can have a more favorable family (Chapter 6), which typically infects the gastroin-
outcome with appropriate care. testinal tract, can give rise to poliomyelitis, a condition
Modes of Virus Infection and Disease 57

characterized by neuronal loss and rapidly progressive used by a particular virus is one key determinant of its
paralysis. HSV-2 too can cause myelitis, as well as men- pattern of dissemination. For example, the rhinoviruses,
ingitis, as noted above. coronaviruses, influenza viruses, and measles virus are all
Before a virus can invade the central nervous system, it transmitted by the respiratory route, and all infect the re-
first must cross the so-called blood-brain barrier. The spiratory tract. However, measles virus is the only one of
physical nature of the blood-brain barrier and the means these viruses that disseminates throughout the body. One
by which certain viruses are able to traverse it are dis- reason why measles virus disseminates whereas these
cussed below. But first, consider that the different clinical other viruses remain in the respiratory tract is that mea-
outcomes that result from virus infections of the central sles virus can bind to several different receptors, including
nervous system result in large part from differences in the CD46, which is present on most cells of the body. The
cell and tissue tropism of different viruses within the cen- consequence of this fact is that serious complications
tral nervous system. Tropism refers to the specific cells from measles can occur inalmost every organ system of
and tissues that a virus infects within its host. For RNA the body.
viruses, tropism is largely (but not entirely) determined Since blood-borne dissemination is the most common
by the cell- and tissue-specific expression of the virus re- means by which viruses spread to infect secondary sites,
ceptor. For DNA viruses, receptor distribution also is a the ability of a particular virus to cross the mucosal barrier
crucial determinant of tropism, but other cellular factors and then invade the underlying tissue and mucosal blood
come into play as well. Regardless, given the vital role of vessels is another determinant of whether a virus can
the central nervous system in the life of a vertebrate, infec- spread systemically. Apropos this discussion, the cells of
tion of this organ system poses serious hazards. the mucosal epithelium and of the vascular endothelium
Rabies virus is much more efficient than HSV at invad- are polarized. That is, tight junctions between these cells
ing the central nervous system. That is, it is much more separate their apical surface from their basolateral sur-
neuroinvasive. However, HSV, like rabies virus, can also face, thereby preventing the free diffusion of membrane
cause fatal disease upon infecting the brain. That is to constituents from one surface to the other. (In the case of
say, HSV and rabies virus are each highly neurovirulent epithelial cells, the apical surface is the exposed free sur-
(Box3.2). Mumps virus is highly neuroinvasive; indeed, it face, and the basal surface contacts underlying cells of the
is more so than even rabies virus. However, mumps virus epithelium or the extracellular matrix. In the case of en-
is not particularly neurovirulent. Thus, neuroinvasiveness dothelial cells, the apical surface faces the lumen of the
and neurovirulence do not necessarily go hand in hand. vessel and is in direct contact with the blood circulation,
Next, we consider the factors that determine whether whereas the basal surface faces the tissue and contacts the
an infection remains local or instead disseminates to be- extracellular matrix.) The physical as well as physiological
come systemic. The distribution of the receptor or receptors separation of the apical from the basolateral surfaces en-
ables each of these surfaces to be specialized to carry out
its particular functions. In that regard, each surface con-
tains receptors unique to that surface.
Box 3.2 If a virus that was replicating in an epithelium was to
assemble and release its progeny from the apical surface
Rabies virus may well be the most lethal virus that infects exclusively, as the rhinoviruses do, then the progeny vi-
humans. The virus typically invades the peripheral rus would end up where the infection began, at the mu-
nervous system following infection by a bite, and mortal- cosal surface. In that event, the infection would remain
ity is nearly 100% once clinical disease occurs. On the
local. On the other hand, if the virus were to assemble
other hand, most of those individuals exposed to rabies
and release its progeny from the basolateral surface, then
virus (i.e., by the bite of an infected animal) do not get
the virus would be in the subepithelial connective tissue,
infected themselves. For example, in a classic study of
positioned to invade and cross the vascular endothe-
this phenomenon, only 2 of 27 people who were bitten
by a rabid wolf in Tehran, Iran, contracted clinical rabies.
lium to invade the blood circulation. Actually, it is a bit
Regardless, if rabies infection should occur, it is almost more complicated than that, since viruses that dissemi-
invariably fatal, whereas untreated HSV infections of the nate in this way generally replicate in the local subepi-
brain are lethal in about 50% of cases. Moreover, while thelial connective tissue, before succeeding in crossing
treatment against rabies is not effective once the virus has the endothelial barrier. Some viruses cross endothelial
invaded the brain, treatment of neural HSV infections barriers by infecting and replicating in the vascular en-
results in a significant drop in fatalities. dothelial cells, whereas others may cross the endothelium
by transcytosis.

Apical release and basolateral release each provide dis-

tinct advantages to a virus. For example, copious apical re-
lease of rhinoviruses from the nasal mucosa facilitates
highly efficient transmission of the virus via droplets from
nasal secretions. Likewise, apical release of poliovirus from
the gastrointestinal mucosa facilitates transmission of that
virus via the fecal-oral route. In contrast, viruses released
from the basal surface have the advantage of moving away
from the mucosal defenses. This ability to disseminate from
the initial site of infection also creates the opportunity for
some viruses to establish a persistent infection at a second-
ary site. The ability to establish persistent infection is an
important viral strategy, since it enables a virus to survive in
small host populations. That is so because the virus has a
home while awaiting new susceptible (i.e., nonimmune)
Figure3.5 Electron micrograph of a sinusoid in fat liver. The en-
individuals to be born into the population (see below). dothelium is extremely attenuated in some areas, where there are fen-
As noted above, some viruses are able to travel via the estrations that give it a sieve-like structure (arrows). From W. Bloom
lymphatic vasculature to nearby draining lymph nodes. and D. W. Fawcett, A Textbook of Histology, 10th ed. (W. B. Saunders
Some viruses do so as free viruses, while others do so by Company, Philadelphia, PA, 1975), with permission.
infecting phagocytes, which then migrate to the nearest
lymph node. Regardless, these viruses then may enter the
blood circulation by passing via the efferent lymph to the the circulation so that they might infect secondary sites.
thoracic duct (Fig. 3.3). Some viruses may undergo fur- Although almost any organ in the body may be infected
ther replication in the lymph nodes before being dissemi- by some virus, individual viruses tend to have preferred
nated via the efferent lymph. exit sites from the circulation, which are based on special
Viruses that are carried by infected lymphocytes or features of the blood-tissue junctions at those sites.
phagocytes are protected from antibodies and comple- The liver, spleen, bone marrow, and adrenal glands
ment in their travels to distant sites. This cell-associated constitute one group of exit sites from the blood that are
form of viremia is a feature of measles virus infection and preferred by viruses. The common feature of the blood-
of several herpesvirus infections. Other viruses, such as tissue junctions at these sites is that in each instance they
HBV, rubella virus, enteroviruses, and flaviviruses (e.g., contain sinusoids, which serve to filter the blood and re-
West Nile virus, St. Louis encephalitis virus, and HCV), move foreign particles. (A sinusoid is a small blood vessel,
circulate as free viruses in the blood. similar to a capillary, but with a discontinuous endothe-
Viremia usually persists until the adaptive immune lium. Some gaps between the sinusoid endothelial cells
system brings the infection under control (Chapter 4). are large enough for blood cells to pass through [Fig. 3.5].)
However, in some cases, the immune system cannot clear To carry out their filtering function, sinusoids are lined
the virus from the host, and the result is a persistent vire- with macrophages, which are a component of the mono-
mia. Persistent viremia is characteristic of HIV and HBV nuclear phagocyte system, previously known as the re-
infections, playing an important role in the pathogenesis ticuloendothelial system. Macrophages of the mononu-
and transmission of each (see below). clear phagocyte system include the peritoneal macrophages
While it is true that a disseminated infection can be of the peritoneal fluid, alveolar macrophages of the lung,
very serious, it would be wrong to assume that localized and microglial cells of the central nervous system. While
infections are necessarily mild. On the contrary, some lo- the major function of the mononuclear phagocyte system
calized infections can be quite severe. For example, rotavi- is to phagocytose microorganisms and foreign substances
ruses can do considerable damage to the intestinal epithe- that are in the blood and various tissues, the anatomy and
lium, leading to severe diarrhea that may be life threatening histology of the blood-tissue junctions in the liver, spleen,
in infants and young children (Chapter 5). Likewise, in- bone marrow, and adrenal glands result in a barrier that is
fluenza virus infections of the respiratory tract epithelium more permeable to viruses than the blood-tissue barrier
produce a serious and life-threatening illness in many in- at other sites.
dividuals (Chapter 12). Viruses in the blood generally can infect reticuloen-
Having accounted for how viruses enter the blood cir- dothelial system macrophages of the liver, which are known
culation, we now need to account for how they exit from as Kupffer cells. The outcome of that infection depends
Modes of Virus Infection and Disease 59

Growth in Kupffer cell

growth in hepatic cell Box 3.3
bile duct excretion
release into blood Viral infection of the pancreas has long been thought
Growth in to be a possible cause of Type I (insulin-dependent)
Kupffer cell diabetes mellitus (Chapter 4). Evidence from seroepide-
miological surveys, and more recently from PCR-based
procedures, implicates enteroviruses, as well as several
other viruses, as etiologic agents for Type I diabetes. For
example, studies show that enterovirus infections ac-
4a 4b 1 company or precede the onset of Type I diabetes in many
Kupffer cell children. Moreover, infection with enteroviruses appears
2 3
to be linked to the induction of autoantibodies against
pancreatic islet cells, as well as to the expression of
Hepatic cells IFN-. Both of these events are connected with islet cell
destruction. Experiments in the mouse model provide
compelling evidence that viruses may cause diabetes in
Inactivation in that system, but there is still no definitive evidence that
Kupffer cell viruses cause diabetes in humans.
Passage through Kupffer cell
growth in hepatic cell
bile duct excretion
release into blood
The kidney glomeruli, the pancreas, the small intestine,
Figure3.6 Types of interactions between viruses and macrophages, and the colon constitute another preferred set of virus exit
exemplified by the Kupffer cells lining a sinusoid in the liver. (1) Mac- sites from the blood. The common feature of the blood-
rophages may fail to phagocytose virions; for example, in Venezuelan
equine encephalitis virus infection, this is an important factor favoring tissue junction at these sites is that the endothelium there
prolonged viremia. (2) Virions may be phagocytosed and destroyed. is fenestrated. That is, there are fenestras, or windows
Because the macrophage system is so efficient, viremia can be main- (more precisely, loose junctions), between the endothelial
tained only if virions enter the blood as fast as they are removed. cells that enable viruses to pass more easily to the underly-
(3) Virions may be phagocytosed and then passively transferred to ad-
jacent cells (hepatocytes in the liver). If, like Rift Valley fever virus or ing tissue. Some viruses, such as HSV, measles virus, and
HBV, the virus replicates in these cells, it can cause clinical hepatitis, yellow fever virus, cross these endothelia as passengers
and the virus produced in the liver can produce a high level of viremia. within macrophages or lymphocytes. These cells enter the
(4) Virions may be phagocytosed by macrophages and then replicate underlying tissue by squeezing between the endothelial
in them. With some viruses, such as lactate dehydrogenase virus in
mice, only macrophages are infected (4a), and progeny virions enhance cells in a process referred to as diapedesis (Box3.3).
the viremia, which reaches an extremely high level. More commonly Virus exit from the circulation is not restricted to those
(4b), as in yellow fever, virus replicates in both macrophages and he- preferred sites enumerated above. For example, viruses
patic cells, producing severe hepatitis. Adapted from D. O. White and also exploit the fact that the blood flow is slowest, and the
F. J. Fenner, Medical Virology, 4th ed. (Academic Press, Inc., San Diego,
CA, 1994), with permission. endothelial barrier is thinnest, within capillaries and
venules. Making use of that fact, a virus can initiate its
passage across the vascular endothelium by binding to a
specific receptor on the apical surface of the endothelium.
The virus might then be transcytosed across the endothe-
on the state of activation of the Kupffer cell, as well as on lium, or it might have to replicate in the endothelium in
the particular virus (Fig. 3.6). Some viruses may be de- order to traverse it. (What factors might determine whether
stroyed by Kupffer cells. Other viruses may replicate in a virus is transported across an endothelium by transcyto-
Kupffer cells, thereby creating an opportunity to infect the sis?) Since virus dissemination across the endothelium by
liver parenchyma. Still other viruses may access the un- either of these possibilities is thought to be inefficient, in-
derlying hepatocytes by transcytosis across the endothe- fection of secondary sites via these routes likely requires
lium. Regardless, viral invasion of the liver may cause se- high concentrations of virus in the blood, a possibility
vere hepatitis, as in the case of yellow fever. In the instances that is favored by extensive replication at the primary site
of HBV and HCV, infection of the liver may be chronic, of infection.
with the possible development of cirrhosis and, less com- As mentioned previously, some viruses may invade the
monly, fatal hepatocellular carcinoma (see below). central nervous system from the blood circulation, and as

just noted, the chances of this happening depend on the junctions, and they are backed by a dense basement mem-
level of the viremia. Moreover, there are two routes by brane that is comprised of an extracellular condensation of
which the viruses might invade the central nervous sys- mucopolysaccharides and proteins. These tight junctions
tem (Fig. 3.7). In the first of these routes, the virus crosses and the underlying basement membrane constitute the
the endothelium of the blood vessels that vascularize the blood-brain barrier. Yet despite this blood-brain barrier,
brain parenchyma. But note that the endothelia of these most viruses that invade the central nervous system do so by
vessels are characterized by unique features to prevent vi- crossing these endothelia. Examples include poliovirus, toga-
ruses from doing just that. The endothelial cells of these viruses, bunyaviruses, and parvoviruses. Some viruses are
capillaries are connected to each other by special tight transported across this endothelium, courtesy of infected
monocytes or lymphocytes. Examples include HIV, measles
virus, mumps virus, and JC virus (a human polyomavirus).
The second viral route from the blood circulation to
Figure3.7 Two routes by which viruses might invade the central
nervous system from the blood circulation. In the first, the virus the brain is via the capillaries in the meninges and choroid
crosses the endothelium of the blood vessels that vascularize the brain plexus. The endothelial cells of these capillaries do not
parenchyma. But note that the tight junctions between the cells of the form tight junctions characteristic of the capillaries that
central nervous system endothelia and the underlying basement vascularize the brain parenchyma. Instead, these endothelia
membrane constitute the so-called blood-brain barrier. Yet despite
this blood-brain barrier, most viruses that invade the central nervous are fenestrated, and their basement membranes are sparse.
system do so by crossing these endothelia. Some viruses do so via This anatomy of the blood-tissue junctions in the meninges
infected monocytes or lymphocytes. The second route by which viruses and choroid plexus enables some viruses to pass from the
pass from the blood circulation to the brain is via the capillaries in the blood circulation into the cerebrospinal fluid. Viruses that
meninges and choroid plexus. The endothelial cells of these capillaries
do not form tight junctions characteristic of the capillaries that serve cross the blood vessels of the meninges or that grow in the
the brain parenchyma. Instead, these endothelia are fenestrated, and epithelium of the choroid plexus may then give rise to men-
their basement membranes are sparse. This anatomy of the blood- ingitis. In these instances, virus particles are found in the
tissue junctions in the meninges and choroid plexus enables some cerebrospinal fluid. Additionally, some viruses in the cere-
viruses to pass from the blood circulation into the cerebrospinal fluid.
Viruses that cross the blood vessels of the meninges or that grow in the brospinal fluid may be able to infect the ependymal cells
epithelium of the choroid plexus may then give rise to meningitis. In that line the ventricles and thereby invade the underlying
these instances, virus particles are found in the cerebrospinal fluid. brain tissue, where they give rise to encephalitis (Fig. 3.7).
Additionally, some viruses in the cerebrospinal fluid may be able to
infect the ependymal cells that line the ventricles and thereby invade
the underlying brain tissue, where they give rise to encephalitis.
The Placenta and the Fetus
Adapted from C. A. Mims et al., Mims Pathogenesis of Infectious The placenta is the organ that exchanges nutrients and
Disease, Academic Press, Orlando, FL, 1995, with permission. waste products between the maternal and the fetal circu-
Direct spread from
latory systems. It is absolutely essential to the survival of
adjacent structures Cerebral all mammalian species. Here, we consider another role for
(e.g., mastoid) blood the placenta, as a protective interface between the mother
and the fetus.
We begin our discussion of the placenta by first noting
Ependyma Brain that it is derived entirely from the embryo. Thus, the pla-
centa, like the fetus, is a foreign tissue within the mothers
body. This fact has given rise to the notion that pregnant
women are in an immunologically compromised state, so
CSF as not to reject either the placenta or the fetus. Another
widely held concept is that newborns are immunologi-
Blood cally ineffectual.
vessel in Neither of the above ideas can be entirely true. Indeed,
choroid Ventricle
plexus if they were true, then it would be most difficult to ac-
CSF count for our survival as a species. That is so because we
Meningeal Nerve are in effect barraged by a battery of potentially lethal in-
blood vessel fectious agents. Yet, remarkably, among viruses only a
handful or so infect the human fetus with any notable
frequency. These include rubella virus, cytomegalovirus
From peripheral
nerve ending or (a herpesvirus), parvovirus B19, and HIV. Other viruses,
nasal mucosa including HSV-1 and HBV, also may be transmitted
Modes of Virus Infection and Disease 61

transplacentally. However, transplacental infections by

these viruses are rarer still.
Box 3.4
Another point to bear in mind is that during birth, the
Since the baby automatically carries the mothers anti
neonate is virtually bathed in a sea of microbes that are
bodies for as long as 2years or more, there is a prob-
present in the mothers genital tract and anorectal sites. lem in determining whether babies of HIV-infected
Consequently, the babys skin, eyes, oropharyngeal mu- mothers who test HIV positive by antibody-based tests
cosa, and possibly upper and lower respiratory tracts are are actually infected, as opposed to just carrying the
exposed to these microbes. One could hardly expect the mothers anti-HIV antibodies. On the same point, in the
baby to survive if it were defenseless against this microbial developed world, mothers of newborns who test HIV
onslaught. Likewise, despite anecdotal reports to the con- positive are advised not to breast-feed their babies. What
trary, objective studies of matched pregnant and non- do you suppose is the reason behind this measure? The
pregnant women have shown no significant difference answer is that the baby might test positive only because it
between them regarding the severity of most infections. carries the mothers antibodies, not because it is actually
Bearing the above points in mind, it is not yet clear infected. If that were the case, then the mother might yet
how the fetus avoids being rejected by the mothers im- transmit the infection to her baby via her breast milk.
mune system, but it is not because the mother is immuno- But what then is the policy in underdeveloped nations?
logically impaired. Quite the contrary is the case, since There, mothers are advised to breast-feed their babies
preexisting immunity in the mother likely acts to impede because of the lack of available health care and adequate
transmission of pathogens to the fetus. (One factor that nutrition.
might act to protect the placenta from maternal immu-
nity is that the fetal trophoblast cells, which constitute the An interesting social commentary is that in much of the
outer layer of the placenta and contact the maternal tissue, underdeveloped world, stigma is attached to a mother
do not express either class I or class II major histocompat- who transmits the infection to her baby, but not to the
ibility molecules [Chapter 4].) man who transmitted the infection to the mother.
The mothers antibodies of course protect the fetus by
protecting the mother from infection. Moreover, mater-
nal antibodies by design (i.e., those of the IgG class
[Chapter 4]) are transferred across the placenta to the fe- ability to cross the placenta, although size does not appear
tus. Thus, maternal antibodies protect the fetus on both to be one of them.
sides of the placenta (Box3.4). There are several potential mechanisms by which vi-
Considering the importance of the placenta to the sur- ruses might cross the placenta, but there is little hard evi-
vival of humans as a species, it is remarkable how little is dence to support any one of them. But before considering
known about how viruses might cross it. One reason for these putative mechanisms, it would be most useful to
our ignorance regarding this issue is that placentas are not have a look at the anatomy of the placenta (Fig. 3.8). Pay
routinely examined in cases of transplacental infections. particular attention to the patterns of the fetal and mater-
Moreover, even in instances where the placenta is exam- nal circulation, noting that the placental villi absorb nu-
ined, infection of the fetus might have occurred without trients from the mothers blood in the intervillous space
any obvious signs of placental involvement. and excrete wastes into that space. On the maternal side of
Since most viruses disseminate via the circulation, the the placenta, blood from branches of the uterine arteries
great majority of transplacental infections of the fetus oc- passes through openings in the basal plate of the placenta.
cur when the mother is viremic. This fact leads us to two From there, maternal blood passes into the intervillous
key questions. First, how is the placenta able to prevent space. Thus, it is reasonable to postulate that viruses can
most of the different viruses that might be in the mothers enter the intervillous space in the mothers blood. A virus
blood from passing to the fetus? Second, what are the spe- might then infect the cells of the villi, thereby enabling the
cial features of those few viruses that do cross the placenta? virus to invade the fetal blood circulation. The villi are
Regarding the second question, we can eliminate size as comprised of trophoblasts, which also line the intervillous
the viral factor, since cytomegalovirus, a large virus, and space. Consequently, trophoblasts would be the major tar-
parvovirus B19, a small virus, both efficiently cross the get cell in this scenario, one that is supported by observa-
placenta. Moreover, EBV, which like cytomegalovirus is a tions that these cells can be infected by some viruses.
herpesvirus, only rarely crosses the placenta, despite the In order for a virus to infect trophoblasts, the virus would
fact that it too is also commonly in the maternal blood. need to express an attachment protein that could recog-
Thus, there must be virus-specific factors that confer the nize a receptor on those cells. Moreover, the trophoblasts

Umbilical cord

Fetal side

ta lm

l blo
m aterna space
of us
ways rvillo
Path ugh inte
ulation thro
Villous tree Fetal circulation Maternal circ
Maternal side in section

Figure3.8 The placentas structure and circulation. The head of maternal blood pressure drives entering blood to-
ward the chorionic plate in fountain-like spurts. As the head of pressure is dissipated, lateral dispersion of blood oc-
curs. Inflowing arterial blood pushes venous blood out into the endometrial veins. The intervillous space is lined by
trophoblasts. The villi absorb nutrients from the maternal blood in the intervillous space and excrete wastes into it.
From W. Bloom and D. W. Fawcett, A Textbook of Histology, 10th ed. (W. B. Saunders Company, Philadelphia, PA,
1975), with permission.

would then need to provide all the factors that the virus fatal AIDS. In contrast, rubella virus and cytomegalovirus are
might need to support its intracellular replication. not fatal to the fetus or to the child after birth. Instead, they
Although the requirements for viral infection of the are teratogenic. That is, they cause abnormal development,
trophoblasts appear rather simple, the placenta still con- which results in birth defects. Indeed, congenital cytomega-
stitutes a formidable defensive barrier against transplacen- lovirus infection is the leading infectious cause of mental
tal virus infections. That might be so, at least in part, be- retardation, deafness, and visual impairment. These viruses
cause even if a virus were to succeed in entering the are each discussed further in their respective chapters. The
placental villi, it still would encounter intervillous defenses, emphasis here is on their interactions with the placenta.
such as the placental macrophages known as Hofbauer Placental involvement in rubella virus infection in
cells. Moreover, even if a virus were to succeed in making its utero has been well documented for more than 40years.
way to the fetus, it would still encounter the fetal defenses, In these instances, damaged foci have been seen in the tro-
which become more imposing as development proceeds. phoblasts that line the villi and the intervillous space.
As noted above, rubella virus, human cytomegalovirus, Moreover, the endothelia of the placental capillaries and
parvovirus B19, and HIV are responsible for the majority larger vessels also show signs of damage. In addition, ru-
of transplacental viral infections. So, what might be the bella virus often can be isolated from the placentas of early
consequences of these infections to the fetus? B19infec- abortion specimens. Nevertheless, by the end of preg-
tions of the fetus usually cause no permanent harm. But nancy, the virus is generally not detected in the placenta,
less than 10% of the time, these infections lead to death of even though it persists in the fetus. What hypothesis might
the fetus and spontaneous abortion. In the case of HIV, you formulate to explain those facts?
children who are infected with this virus in utero and who Cytomegalovirus reaches the fetus in about 40% of pri-
are not given antiretroviral therapy will eventually develop mary maternal infections, and this virus is present in as
Modes of Virus Infection and Disease 63

many as 2.3% of all newborns. (A primary infection is the

initial introduction of an infectious agent into the body.
Box 3.5
Cytomegalovirus gives rise to latent infections that spon-
As is discussed at greater length in Chapter 21, the AIDS
taneously reactivate [see below]. A mother initially infected
epidemic has compelled individuals, and society as a
with the virus during pregnancy will not have antibodies in whole, to confront old attitudes and prejudices and to
place to control her infection or to pass on to the fetus. In make difficult moral judgments. One such complicated
contrast, if a mothers latent infection reactivates during moral judgment concerns making choices in the admin-
pregnancy, she will have antibodies in place to control her istration of AIDS therapies, particularly in underdevel-
infection and to pass on to the fetus.) Evidence for placental oped countries, in the face of high costs. Consider that
involvement in cytomegalovirus infections in utero comes antiretroviral therapies administered to the mother dur-
primarily from the analysis of abortion specimens. In these ing pregnancy significantly reduce the incidence of HIV
specimens, virus can be detected in villus trophoblasts and transmission to the fetus. At first glance, the decision
in the vascular endothelium. In addition, explants of fetal here would appear to be obvious. Tragically, however, this
trophoblasts can be infected with cytomegalovirus invitro. practice also leads to many abandoned orphans, many of
As noted previously, most transplacental infections in utero whom are HIV infected despite treatment and who will
are transmitted from the maternal blood circulation. But, in eventually themselves succumb to AIDS. In the coun-
the case of cytomegalovirus, virus that is excreted from the tries that are hardest hit by HIV, including Zimbabwe,
cervix of infected women (after a recurrence) also can get to Botswana, Namibia, Swaziland, and South Africa, one in
the fetus by ascending up the genital tract to the amnion. three children are presently orphans of parents who suc-
About 30% of women who are infected with parvovi- cumbed to AIDS. In point of fact, 50% of new mothers
rus B19 during pregnancy pass that virus on to the fetus. in South Africa likely will die of AIDS. Bearing these facts
Thus, parvovirus B19, like cytomegalovirus, appears to be in mind, it is not universally agreed that saving an infant
rather efficient at crossing the placenta. Placental involve- (perhaps only temporarily) of an HIV-infected mother is
ment in these fetal infections is evident from the consider- the wisest use of the limited funds that might be available
for all worthy public health measures in resource-limited,
able damage to the placental vascular endothelium.
underdeveloped countries. What do you think?
The consequence of HIV infection in utero is enor-
mous. It is estimated that during 2002 and 2003, some
1.6 million children worldwide were born infected with
HIV. Although this virus also may be transmitted to a
newborn via the mothers milk, experimental evidence
shows that over 90% of pediatric AIDS patients acquired that the latter may display. We begin by enumerating the
the virus in utero from an HIV-infected mother or during defining characteristics of acute infections. Viruses are
passage through the birth canal. In most instances of HIV rapidly produced during the course of an acute infection,
infection in utero, the fetus is infected sometime after the and symptoms appear suddenly if there is an associated
first trimester, but infection can occur as early as the illness. Acute infections terminate in either of two possible
eighth week of gestation. The likelihood that the fetus will outcomes. In one outcome, the infection is rapidly cleared
be infected by an untreated infected mother is estimated by host immune defense mechanisms, and the patient re-
to be as high as 30 to 50% (Box3.5). covers and is immune to the specific virus. The alternative
The mechanism of HIV transmission in utero is not well outcome is for the host to quickly succumb to the infec-
understood, but HIV type 1 is detected in the placenta invivo, tion. In either case, the duration of the episode is generally
and it can replicate in placental explants invitro. Indeed, pla- less than 2weeks.
cental trophoblasts are susceptible to HIV infection invitro, The alternative courses that a virus infection might fol-
although about 10-fold less so than CD4 T cells, which are low are referred to collectively as persistent infections, in
the primary target cells of HIV. In addition, there are research which the virus may persist for years, perhaps even for the
reports that HIV may also cross the polarized trophoblast normal life span of the infected individual. Such persis-
cell layer by transcytosis. This process may occur indepen- tent infections are categorized as latent, chronic, or slow,
dently of, or concurrently with, virus replication. depending on the pattern of virus production and actual
disease episodes (Fig. 3.9). The defining features of these
different sorts of infections are noted below, and several
ACUTE VERSUS PERSISTENT INFECTIONS examples are considered in each instance. Some specific
This section considers the key distinctions between acute diseases are discussed in more detail in the chapters per-
and persistent virus infections and the variety of patterns taining to the particular virus families.

Patterns of virus infection

Acute infections Rhinovirus
Influenza virus
Measles virus

Slow infection Measles (SSPE)

Human T-cell leukemia virus type 1
Post-polio syndrome

Hepatitis B
Chronic infection: early disease Hepatitis C
JC virus, progressive multifocal
Figure3.9 Diagram of the general Papillomaviruses
patterns of a viral infection, with HIV
several selected examples indicated. Chronic infection: late disease
The time scale is arbitrary. Areas in
orange indicate the presence of
virus and disease. Areas in tan indi-
cate the presence of virus but no
disease. When under a solid line, Latent infection Herpes simplex virus type 1
virus is readily demonstrable. Varicella-zoster
When under a dashed line, virus is
not readily demonstrable. Some of
the viruses named as examples may
give rise to infections that have the
characteristics of more than one Transmissible spongiform encephalopathies Scrapie
category. Modified from D. O. Kuru
Creutzfeldt-Jakob disease
White and F. J. Fenner, Medical BSE
Virology, 4th ed. (Academic Press,
Inc., San Diego, CA, 1994), with
permission. Time (years)

Acute Infections would be useful to first note that the mammalian immune
Viruses such as the rhinoviruses and influenza viruses, system is comprised of two branches: the innate immune
which remain localized at the primary site of infection, system and the adaptive immune system (Chapter 4). The
generally give rise to acute infections. Although some lo- innate immune system is ready to respond immediately
calized acute infections (for example, those involving in- and nonspecifically to a broad range of invaders. In con-
fluenza viruses) can be lethal, localized acute infections trast, the adaptive immune system needs to be activated
generally are less life threatening than disseminated infec- to tailor a specific response to each particular invader.
tions, and they are commonly resolved with the elimina- Whereas adaptive immunity is required to resolve infec-
tion of the virus from the body. tions by most viruses, some sources state that the innate
Because those viruses that give rise to acute infections immune system is often sufficient to resolve rhinovirus
generally infect cells that make up the mucosal surfaces, infections, since these infections resolve within just a few
they usually are efficiently transmitted from one individ- days, that is, before the adaptive immune system might be
ual to another. This is so despite the fact that these infec- fully activated. Whether or not a rhinovirus infection
tions are usually quickly resolved. As a consequence of might actually be resolved by the innate immune system
their high transmission rates, viruses that give rise to acute alone is discussed below and in Chapter 4. Regardless, the
infections tend to be associated with widespread epidem- early resolution of rhinovirus infections poses no particu-
ics that might affect millions of individuals. Influenza is a lar obstacle to the success of these viruses, since they rep-
good case in point. (See Box11.14 for the other character- licate very efficiently during that brief interval of only
istics of viruses that give rise to periodic epidemics.) several days. Moreover, the inflammatory response, which
The rhinoviruses provide a good example of a virus in reality is a defensive stratagem of the innate immune
group that is well adapted to exploit the acute-infection system, produces the runny nose, sneezing, and cough
lifestyle. To better appreciate how rhinoviruses do so, it reflex that facilitate the efficient transmission of rhinoviruses
Modes of Virus Infection and Disease 65

from one individual to another. Furthermore, if the adap- disseminating virus that gives rise to acute infections. Like
tive immune system is not fully activated, then there may influenza viruses and the rhinoviruses, measles virus
be little acquired immunity against the same rhinovirus spreads by the respiratory route. However, the course of a
serotype. (Serotypes are groups of viruses that are defined measles virus infection is quite different from that of the
by their interactions with antibodies. Neutralizing anti- two classic respiratory viruses. The initial measles virus
bodies against one serotype do not neutralize another se- infection of the respiratory mucosa produces little trans-
rotype. As noted below, there are more than 100 distinct missible virus. The reason is that innate immunity is ex-
rhinovirus serotypes that circulate in humans.) tremely effective at controlling the initial measles virus
Considering the unique capability of the lymphocytes infection at that site. Thus, measles virus would likely not
of the adaptive immune system to target and destroy do very well as a respiratory virus unless it had some ad-
virus-infected cells (Chapter 4), many virologists and im- ditional stratagem up its viral sleeve. Indeed, measles vi-
munologists would not expect that a virus infection could rus has adapted to infect dendritic cells at the respiratory
be completely resolved in the absence of adaptive immu- mucosa. (Dendritic cells, mentioned above, are discussed at
nity. Also arguing against the premise that rhinovirus in- length in Chapter 4. We note here that they serve as sentry
fections might be resolved by innate immunity only is the cells in the tissues, where they phagocytose pathogens and
fact that increased numbers of T lymphocytes (which me- their antigens, which they then deliver to lymph nodes. In
diate the adaptive immune response against virus-infected the lymph nodes, they present these antigens to the effector
cells) are found in nasal secretions only 3 to 4days follow- cells of adaptive immunity, thereby activating an adaptive
ing experimental inoculation with rhinoviruses. immune response.) Measles virus not only survives in den-
But if adaptive immune responses do play a role in the dritic cells but also replicates in them. Thus, when a measles
resolution of rhinovirus infections, then infection should virus-infected dendritic cell arrives at a lymph node, it turns
also result in specific immunity against reinfection by the the lymph node into a secondary site of infection. But mea-
same virus (Chapter 4). Why then do we have rhinovirus sles virus is yet more diabolical; infection of lymph nodes
common colds throughout our lifetimes, sometimes as temporarily impairs the function of the T cells and B cells
many as several per year? First, there are more than 100 that mediate adaptive immunity, thereby increasing the ef-
antigenically distinct rhinovirus serotypes circulating in ficiency of the secondary infection.
humans. Consequently, there are always some serotypes The measles virus subterfuge described above results in
against which a person may have little, if any, acquired im- a massive secondary viremia that enables the virus to in-
munity. Second, human rhinoviruses use several different fect other tissues, as well as to reinfect the respiratory mu-
strategies to evade neutralization by antibodies, as dis- cosa. The secondary attack on the respiratory mucosa is
cussed in detail in Chapters 4 and 6. initiated by many more viruses than initiated the primary
Influenza viruses, like rhinoviruses, cause only acute infection. Consequently, sufficiently large numbers of vi-
infections, and in the case of influenza, the adaptive im- rus are produced during the reinfection of that mucosa to
mune system is clearly needed to resolve the infections. enable very efficient transmission of the virus to new indi-
Thus, influenza virus may have several days more than the viduals. Indeed, measles virus is one of the most conta-
rhinoviruses to transmit the infection from one individ- gious viruses in humans, infecting about 40million indi-
ual to another. The first new influenza viruses appear in viduals worldwide per annum. One to two million of
the respiratory mucosa, the first day after infection. Virus those infections are fatal. This complex course of events
levels peak at about the fourth day, when the adaptive im- during measles virus infections explains why there usually
mune system springs into action. Virus levels then begin are no symptoms until 10days or so. (A recent new study
to decline by day 6, but the patient may continue to be implies a different scenario of measles virus infection, in
infectious for several days more. One distinction between which lung lymphocytes are the first cells to be infected.
a rhinovirus common cold and influenza is that the pa- These primarily infected lymphocytes then transport the
tient may feel well soon after the rhinovirus infection has virus to the basolateral surface of the airway epithelium,
cleared, while the influenza patient may continue to feel ill which then sustains measles virus shedding into the air-
for days, even weeks afterwards, due primarily to lingering way lumen, a step that is crucial for transmission of the
effects of the immune responses (Chapter 4). virus. Importantly, primary infection of the lymphocytes
Recalling that rhinoviruses and influenza viruses re- was found to be sufficient to give rise to viremia and clinical
main at the initial site of infection, note that some dis- measles. Experimental details are recounted in Box11.7.)
seminating viruses also give rise to acute infections. In In approximately 1 of every 300,000 measles infec-
those instances, the infections at both the primary and the tions, the virus persists in the brain of the infected indi-
secondary sites are acute. Measles virus is an example of a vidual for years, eventually producing an invariably fatal

neurodegenerative disease called subacute sclerosing particular virus families. The second need also has im-
panencephalitis (SSPE). Persistent virus infections are mune implications, since the death of infected cells is one
discussed at some length below. We mention SSPE here of the main triggers of innate immunity, which in turn
because this late complication of measles virus infections initiates adaptive immune responses (Chapter 4).
demonstrates that some viruses that are generally associ- It is convenient to categorize persistent virus infections
ated with acute infections also may give rise to persistent as slow, chronic, or latent, based on the pattern of virus
infections. See also Slow Infections, below, for more in- production and the emergence of symptoms (Fig. 3.9). Yet
formation on SSPE. some persistent infections show features of more than one
of these categories and cannot be pigeonholed neatly into
Persistent Infections any one of them. Thus, note that other authors may use
Persistent virus infections usually begin as typical acute examples that here illustrate one category of persistent in-
infections but then develop into one of the recognized fection as an example of another.
categories of long-term persistent infections (i.e., latent,
chronic, and slow). Some persistent infections are associ- Slow Infections
ated with very serious human illness, as evidenced by Slow virus infections may begin as fulminating acute in-
such diseases as AIDS, resulting from infection with HIV; fections (i.e., ones where illness comes on suddenly, with
cirrhosis and hepatocellular carcinoma, resulting from severe symptoms of short duration). However, a defining
chronic infections with HBV and HCV; and cervical car- characteristic of slow infections is that in virtually all in-
cinoma, associated with the human papillomaviruses. stances disease progresses slowly and relentlessly towards
From the point of view of the virus, persistent infec- death. Note that slow more appropriately refers to the
tion is an important strategy that enables it to maintain time course of the disease per se, not the rate at which the
itself in the host population, particularly in small popula- virus replicates. Virus replication may very well be rapid,
tions. This is a very important point. Virus persistence during the acute phase at least. Afterwards, however, rep-
provides time for new susceptible individuals to be born lication may occur at a low level. For this reason, some
or migrate into the host population. These new members authors prefer the term smoldering infection when re-
of the host population then might themselves become ferring to this class of infections. Another characteristic of
carriers of the virus. (Measles virus, which gives rise only slow infections is that virus generally can be recovered
to acute infections, requires a host population of at least from the experimental animal or patient during the long
500,000individuals in order to ensure a sufficient quan- preclinical phase, as well as after symptoms have emerged.
tity of new susceptible individuals to sustain the virus in These infections also have been called acute infections
the population [Chapter 11]. Because populations of such with rare late complications. The continuous low levels
size did not exist in preagricultural times, measles virus of virus replication characteristic of slow virus infections
could not have been sustained in the human population tend to continually stimulate the T-cell population, leav-
until relatively recently in human history. But humans are ing little time for its recovery. This may lead to impaired
the only host for measles virus. This seeming paradox is T-cell function, thereby facilitating viral persistence.
explained in Chapter 11. In contrast to the acute measles SSPE exemplifies a slow virus infection (Fig. 3.9). As
virus, persistent viruses, such as the herpesviruses, have noted above, measles virus persists in the brain of approx-
been coevolving with their hosts for millions of years imately 1 of every 300,000individuals who are naturally
[Box18.11].) infected with measles virus. After an incubation period of
For a virus to establish a long-term persistent infection, 1 to 7years, the virus may give rise to SSPE, which is in-
it must avoid being cleared by the immune system of the variably fatal. This disease is discussed in greater detail in
host. Moreover, in instances where the virus is cytocidal, Chapter 11. For now, note that SSPE patients display little,
the virus must also be able to modulate expression of if any, measles virus in the brain. However, they have ex-
those viral genes that sustain lytic infection, so as not to ceptionally high titers of neutralizing anti-measles virus
kill the host cell. The first need might be met in part by the antibodies in the cerebrospinal fluid. Together, these facts
persistence of the virus in immunologically privileged imply an ongoing low level of virus production in the
sites, such as the central nervous system. Perhaps this ex- brain (i.e., a smoldering infection). This slow replication
plains why so many persistent virus infections affect the and spread of measles virus in the brain and the high
central nervous system. Additionally, persistent viruses antibody titers in the cerebrospinal fluid are defining
generally have evolved other rather clever means for evad- characteristics of SSPE. The virus enters the brain via
ing host immunity. These strategies are reviewed in Chap- lymphocytes, which it infects during the primary or sec-
ter 4 and are discussed further in chapters dealing with the ondary viremia (see above). Much remains to be learned
Modes of Virus Infection and Disease 67

concerning the viral and host factors that underlie virus cell, localized at the synapse, which contain the viral RNA
persistence and pathogenesis in SSPE. genome and the precursor polyprotein (Gag) for the
Maedi/visna (initially recognized in the 1930s, when it viral reverse transcriptase (Chapter 20). Moreover, cell-
appeared in Icelandic sheep) was in actuality the first slow to-cell contact induces the rearrangement of the cy-
virus disease to be characterized. In fact, the term slow toskeleton beneath the complexes, which implies that
infection was originally coined to describe this condition. microtubules are involved in transporting RNA/Gag
Maedi/visna is caused by a lentivirus, a subgroup of the complexes to the synapse and into the recipient cell. These
retroviruses that also includes HIV. It is a demyelinating experimental findings are consistent with the apparent fact
disease of sheep, with an incubation period of about 2 to that lymphocytes that are naturally infected with HTLV-1
3years, which progresses slowly but inevitably to death. do not readily produce free enveloped extracellular virus
Virus is found intermittently in the cerebrospinal fluid and particles. Note that HIV too can be transmitted by direct
blood, and neutralizing antibodies are present, although cell-to-cell contact via a virological synapse (Chapter 21).
they obviously do not eliminate the virus. The virus is not The ensuing replication cycle of HTLV-1 is most un-
cytopathic (i.e., virus replication and release per se do not usual, even for a retrovirus. And the atypical HTLV-1 rep-
lead to cell death), and disease may be due to an antigen- lication cycle is important in the current context because
antibody response on the surface of infected glial cells. it explains why, in contrast to other infections that lead to
Another retrovirus, the human T-cell leukemia virus slow virus diseases, HTLV-1 infections do not begin as
type 1 (HTLV-1), is asymptomatic in more than 95% of fulminating acute infections. First note that a key step in
infected individuals. Still, it is responsible for several rare, the replication of all retroviruses, including HTLV-1, is for
late complications. One is adult T-cell leukemia, which the plus-strand RNA genome to be reverse transcribed
has an incidence of about 2% in infected individuals. An- by the viral reverse transcriptase into a double-stranded
other is HTLV-1-mediated tropical spastic paraparesis, DNA genome, which is then integrated by covalent bonds
again with an incidence of about 2%. This is a paralytic into the cellular genome (Chapter 20). Importantly, re-
disease, characterized by demyelination of long neurons verse transcription per se does not replicate the viral ge-
in the spinal cord. This pathology is believed to result nome. Instead, it merely exchanges the single-stranded
from an autoimmune attack on the myelin, brought on by RNA genome for one comprised of double-stranded DNA.
the persistent HTLV-1infection. Each of these HTLV-1- Most retroviruses then amplify their genomes by transcrib-
mediated diseases emerges 30 to 50years postinfection! ing the integrated proviral DNA genome, as mediated by
HTLV-1 is found almost exclusively in association with the cellular RNA polymerase II. (The term provirus
CD4 T cells in infected individuals. That is to say, there is usually refers to the integrated DNA form of a retrovirus
little, if any, free virus that can be detected in these pa- genome.) In contrast, HTLV-1 replication and persistence
tients. Consequently, HTLV-1 is believed to be transmit- occur mainly as follows. Like other retroviruses, HTLV-1
ted by means of these cells, which are present in semen, generates a DNA copy of its genome, which is then inte-
blood, and breast milk. Accordingly, routes of transmission grated into the host cell genome. When the cell divides,
are believed to be via sexual contact, through exposure to the integrated provirus genome is replicated along with
contaminated blood (either through blood transfusion or the cellular DNA and is passed in the integrated state to
sharing of contaminated needles), and from mother to each of the daughter cells. Since free virus is very ineffi-
child via breast-feeding. Surprisingly, perhaps, free HTLV-1 cient at infecting new cells, this is the principal means by
is actually very inefficient at initiating infections. which the HTLV-1 infection is amplified. And, as noted
Here now is an interesting point. Since in each of the above, the infection is transmitted to new individuals via
above modes of HTLV-1 transmission infected cells, rather these infected cells. (This means of viral replication might
than free virus, must be passed from an infected individual, be referred to as clonal proliferation. What experimental
how does the virus actually infect the recipients cells? Re- evidence might support this mechanism? Most importantly,
markably, when an HTLV-1-infected lymphocyte contacts treatment of HTLV-1-mediated tropical spastic parapare-
an uninfected lymphocyte, a virological synapse forms sis patients with reverse transcriptase inhibitors, which
between the two cells. Both cell-to-cell contact (involving are routinely used in antiretroviral therapy [Chapter 4],
the cell surface proteins ICAM-1 and LFA-1 [Chapter 4]) does not change the proviral loads in these patients. This
and expression of the HTLV-1 Tax protein are necessary would be expected if free virus does not contribute to the
to induce the synapse. One study (Igakura etal., 2003) re- proviral burden.)
ported that this contact between an HTLV-1-infected HTLV-1-infected cells are not killed by these events.
lymphocyte and an uninfected lymphocyte induces the This is an important point, since the nonlytic nature of
formation of macromolecular complexes in the infected the virus is a factor enabling it to persist. Another point to

note is that the proviral DNA can be integrated into the liver. However, despite infecting the same target organ and
host cell DNA only when the host cell is replicating. Thus, causing similar clinical outcomes, there nevertheless are
the virus remains in an as yet poorly defined kind of latent marked differences in the patterns of chronic infection of
state in an infected cell until such time as the cell might these two distinct viruses. JC virus, a human polyomavi-
replicate, which, in the case of a CD4 T cell, occurs when rus that causes a demyelinating disease in the brain, and
the cell is presented with its cognate antigen (Chapter 4). human papillomaviruses, which cause benign warts and
As we will see when we consider the features of chronic cervical carcinoma, demonstrate still other patterns of
and latent infections, HTLV-1-associated infections display chronic virus infection. Finally, HIV, which causes AIDS
features of these other categories of persistent infections, (a disease that is particularly difficult to place in any one
as well as of slow infections. category), is briefly considered here among the chronic
One last thought regarding HTLV-1 is that the mecha- infections.
nism by which this leukemia virus induces neoplasia is HBV causes the most important chronic virus infection
different from that of other oncogenic retroviruses. Un- in humans. Actually, HBV is one of the most important hu-
like many of these other retroviruses, HTLV-1 does not man pathogens of any kind, as there are about 500million
harbor a viral homolog of a cellular proto-oncogene. (An chronic human carriers worldwide and over 1million fa-
oncogene is a gene that can cause a cell to become malig- talities each year due to HBV-associated liver disease. In
nant. Retroviral oncogenes are generally derived from nor- some areas of the world, including Southeast Asia and
mal cellular counterparts that have been incorporated into southern Africa, as many as 50% of all individuals are in-
viral genomes and altered so they malfunction when ex- fected. (HBV, a member of the hepadnavirus family [Chap-
pressed in an infected cell [Chapter 20].) They do not play ter 22], is ostensibly a double-stranded DNA virus. But in
a role in virus replication. In contrast, the primary neoplas- actuality, its genome is partially double stranded and par-
tic activity of HTLV-1maps to a viral gene that encodes its tially single stranded. While this may seem a bit bizarre, it
Tax protein, which plays a key role in virus replication and is not nearly as bizarre as the overall replication strategy of
which has multiple effects on intracellular signaling path- this virus. In brief, the HBV replication strategy involves an
ways, cell cycle progression, and chromosome stability. RNA intermediate transcribed from the viral genome,
Postpolio syndrome, a late sequela (i.e., a disease that which is packaged into empty virus capsids, where it is then
is caused by a preceding disease or injury in the same in- reverse transcribed by a virus-encoded reverse transcriptase
dividual) of poliomyelitis (Chapter 6), illustrates another activity, thereby generating the DNA genome.)
variation of a slow virus infection. This condition may oc- Several features of HBV are thought to facilitate estab-
cur 30 to 40years after the original poliovirus infection, in lishment of a chronic infection, two of which are noted
as many as 80% of the original victims. Affected individu- here. First, HBV is not cytocidal, and second, it does not
als suffer a further deterioration of the originally affected induce the early production of interferons (IFNs). (The
muscles. In this condition, the virus cannot be detected IFNs are briefly introduced in Box3.6 and are discussed at
and may very well not even be present. It is believed that length in Chapter 4 and elsewhere in the text as well.)
the syndrome results from a later loss of neurons in the Since tissue damage and IFN production together gener-
initially infected nerves, due to the greater burden borne ally induce an inflammatory response in the host, which
by them over the years. usually is necessary to activate a specific adaptive immune
response against a virus (Chapter 4), these features of
Chronic Infections HBV enable it to steer clear of alarming the immune sys-
Chronic virus infections are generally characterized by con- tem, while it establishes persistence.
tinuing high levels of virus replication and viremia. However, Eventually, HBV does activate the adaptive immune
disease may be absent, or it may be chronic, or it may develop system, although it is not entirely clear how it does so.
late in some instances (Fig. 3.9). In any case, and in contrast Maybe it is because in due course a sufficient number of
to slow virus infections, chronic infections do not lead inexo- infected cells are killed. Regardless, B lymphocytes begin
rably to death. (Chronic infections with some viruses may, in to produce antibodies against the virus, and virus-specific
rare instances, lead to fatal neoplasia. For example, chronic cytotoxic T cells are activated as well. The adaptive immune
HBV infections may progress to fatal hepatocellular carci- response then resolves the infection in more than 90% of
noma [see below]. Nevertheless, hepatocellular carcinoma is adult cases. Moreover, these individuals experience few, if
rare, even among HBV carriers. Still, it is about 100-fold any, symptoms, and upon recovering they have lifelong
more prevalent in these individuals than in controls.) protective immunity. In the remaining infected adults,
HBV and HCV are two unrelated viruses, each of which there is sufficient liver damage to cause the characteristic
gives rise to medically important chronic infections of the nausea, jaundice, and liver pain associated with hepatitis.
Modes of Virus Infection and Disease 69

Box 3.6 Now here is a bit of a puzzler. Since HBV is itself non-
cytocidal, what processes might underlie the liver damage
and hepatitis symptoms associated with HBV infections?
IFNs are a group of antiviral proteins that are expressed
In point of fact, the inflammatory response of the innate
by cells in response to a viral infection (Chapter 4). There
are three classes of IFNs, IFN-, IFN-, and IFN-. IFN-
immune system and cytotoxic T cells of the adaptive im-
and IFN- are referred to collectively as IFN-/ or mune system are together responsible for HBV pathogen-
type I IFNs. They are synthesized by, and secreted from, esis. Indeed, the immunopathological responses that are
almost any type of virus-infected cell. They bring about induced by some infections are the inadvertent price we
an antiviral state in neighboring cells by transmitting sometimes must pay to enjoy the protection afforded by
signals mediated by IFN receptors on those cells which our otherwise quite remarkable immune systems (Chap-
upregulate the expression of IFN-responsive genes. IFN- ter 4). Regardless, the conclusion that immunopathology
is produced by certain lymphocytes after they have been underlies the liver damage in HBV-infected individuals is
exposed to a foreign antigen, and it is capable of enhanc- consistent with the observation that infections in patients
ing the efficacy of the immune system. may change over from an immunotolerant phase to an
immunoactive one. (An individual is said to be immuno-
Since the genes for IFN-/ are expressed in response tolerant when he or she is nonreactive to a particular an-
to a virus infection, note that HBV does not seem to tigen. Generally, we are immunotolerant of self antigens.)
affect the expression of any liver cell genes, including the The immunotolerant phase is characterized by higher vi-
IFN-responsive genes, at least as indicated within the rus titers but less-severe liver disease, whereas the immu-
sensitivity of microarray analysis. On the other hand, noactive phase is characterized by low virus titers and
experimental results from a chimpanzee model suggest more-severe hepatitis. Note that the immunoactive, low-
that IFN- may have a role in regulating HBV infections. replication phase can last for the lifetime of the patient
In that model, most of the virus is cleared before the de- (Box3.7).
velopment of an adaptive immune response, and IFN- It is important to distinguish HBV infections in adults
may be a key factor in that clearance (see the text). Much
from infections by that virus in the very young. This dis-
remains to be learned about this intriguing infection.
tinction is clinically important and also important with
respect to the natural history of the virus. We begin by
comparing routes of transmission in these two instances.
Important routes of HBV transmission between adults
include the sharing of syringes among intravenous drug
Interestingly, although recovered individuals are dis- users, transfusions of contaminated blood, acupuncture,
ease free and immune to reinfection with HBV, they yet body piercing, tattooing, and sexually via blood and saliva.
may produce trace amounts of the virus, with the low- Of these routes of transmission between adults, only the
level infection regulated by the adaptive immune response. last might be considered a natural route. With that in mind,
Actually, it is not clear whether or not HBV is ever com- the most important natural route of HBV transmission is
pletely eliminated from infected individuals. For example,
clinically recovered patients who have later been immu-
nosuppressed by cancer chemotherapy are known to have
experienced reactivation of their hepatitis B infections.
T lymphocytes were shown recently to have a crucial Box 3.7
role in the control of HBV infections. This was demon-
Other evidence that immunopathology underlies HBV
strated in chimpanzees, in which experimental depletion of
pathogenesis includes experimental results obtained in a
either helper T cells or cytotoxic T cells prevented virus clear-
mouse model. Transgenic mice, which express the hepa-
ance and clinical recovery. In addition, a still-mysterious titis B surface antigens, display no symptoms whatso-
noncytolytic antiviral mechanism also is at work, as im- ever. Moreover, these mice are immunotolerant of those
plied by the observation that most of the virus can be virus antigens. That is, they do not recognize those virus
cleared from the liver and blood of infected chimpanzees antigens as foreign. (Why do you think that might be so?
before there is any detectable T-cell infiltration of the liver See Chapter 4.) However, when these mice are injected
or liver injury. Studies using a transgenic mouse model with hepatitis B-specific cytotoxic T cells, they develop
suggest that this mysterious factor might be gamma in- pathologies that strongly resemble pathogenic human
terferon (IFN-) released by activated cytotoxic T cells HBV infections.

actually from a chronically infected mother to her child at cancers contain the viral genome, strongly implying that
birth. When virus levels are sufficiently high in a chroni- expression of a viral function in the neoplastic cell in some
cally infected mother, then there may be a high rate of way promotes cancer development.
vertical transmission (i.e., from mother to offspring). To As stated above, HBV is a member of the hepadnavirus
reiterate, the most important natural route of HBV trans- family of partially double-stranded DNA viruses, which
mission is vertical transmission from mother to offspring. replicate via the rather astonishing reverse transcription
HBV infection during adulthood typically does not re- of an RNA intermediate (Chapter 22). In contrast, HCV
sult in chronic hepatitis but instead leads to protective im- is a member of the flavivirus family of conventional,
munity. However, in about 3% of individuals infected as positive-strand RNA viruses (Chapter 7). But despite the
adults, the virus prevails against the immune system to fact that HBV and HCV are unrelated, each gives rise to
establish a long-term, chronic infection that can persist medically important chronic infections of the liver, which
for life. Those individuals infected as adults who become in each instance may lead to chronic hepatitis, cirrhosis,
chronically infected might have mounted an inefficient and hepatocellular carcinoma. These facts illustrate the
immune response. In addition, the clever virus itself con- point that biologically unrelated viruses may infect the
tributes to the inefficiency of the immune response by same target organ and cause similar clinical outcomes.
generating noninfectious empty particles, which are lack- HCV infects about 170million people worldwide. Still,
ing the viral genome. These particles may exceed the despite the fact that HCV affects so many individuals, as-
number of infectious particles by a factor of 104 to 106 and pects of the virus-host interaction that might underlie
are believed to serve as decoys by absorbing most or all of chronic HCV infection in vivo are not yet understood.
the neutralizing antiviral antibodies. This is so largely because a system for the study of HCV in
In contrast to the majority of individuals infected as cell culture has not yet been developed. Also, little is
adults, virtually all infected newborns go on to become known regarding the natural routes of HCV transmission.
chronic carriers themselves, perhaps because of their im- Like in the case of HBV, vertical transmission from mother
mature (but hardly inert; see above) immune systems, to offspring at birth is the only well-documented natural
although other explanations are plausible (e.g., route of route of HCV transmission. There is evidence that the vi-
infection or dosage). Moreover, individuals infected as rus also may be transmitted sexually, but that appears to
newborns suffer less tissue damage and have milder symp- be a rather inefficient route. As expected, HCV is very
toms than individuals contracting the infection as adults, efficiently transmitted via infected blood.
again possibly because of a less effective immune response. In contrast to HBV, which establishes a chronic infec-
The infected newborns, as chronically infected carriers, will tion in only about 3% of individuals infected as adults,
one day be able to efficiently pass the infection on to their HCV establishes chronic infections in most (i.e., 60 to
offspring, thereby accounting for the fact that vertical trans- 80%) of those infected as adults. Moreover, most individ-
mission is the most important natural route of hepatitis B uals infected with HCV as adults eventually develop
transmission. Indeed, HBV has evolved a very efficient chronic hepatitis. Also, the development of chronic hepa-
means to ensure its maintenance in the host population. titis actually correlates with a 100- to 1,000-fold drop in
The chronic HBV infection may be asymptomatic, or it the virus titer. (How might you explain the correlation be-
may be associated with chronic active hepatitis. The latter tween the drop in the HCV titer and the emergence of
condition may progress to potentially fatal cirrhosis (a chronic hepatitis? See above with respect to hepatitis B.)
chronic progressive disease of the liver characterized by HCV must evade innate immunity at early times in the
the replacement of healthy liver tissue with scar tissue) infection, in order to establish persistent infection. In that
and, more rarely, to hepatocellular carcinoma, a usually regard, the virus evades the effects of the IFNs, which are
fatal malignancy with a latency period of between 9 and among the effectors of innate immunity (Box3.6). Recall
35years. that HBV avoids the effects of the antiviral type I IFNs by
HBV does not express an oncogene (see above). In- not inducing their expression in the first place. In contrast,
stead, the development of hepatocellular carcinoma is HCV does upregulate the expression of many liver cell genes,
believed to result from the extensive cell proliferation that including those involved in the production of type I IFNs.
takes place in the liver, which occurs for the purpose of Nevertheless, HCV avoids the effects of these IFNs by pro-
replacing the cells that are continually destroyed by the ducing two proteins, E2 and NS5A, each of which inhibits
immunopathology associated with the infection. This ex- the activity of RNA-activated protein kinase (PKR), a me-
tensive cell proliferation may increase the risk that a cell diator of the IFN antiviral effect (Chapter 4). E2 acts as a
will acquire cancer-causing mutations. Yet this cannot decoy substrate for PKR, whereas NS5A forms heterodim-
be the whole story, since all hepatitis B-associated liver ers with PKR to inhibit its activity.
Modes of Virus Infection and Disease 71

Since the most important natural route of HCV trans- liver damage transpires over the course of 20 to 30years to
mission may be from an infected mother to her child at lead to cirrhosis. Within 10 more years, about 10% of
birth, an infected female child must carry the virus until these affected individuals eventually develop hepatocellu-
she reaches maturity if she is to be able to transmit the lar carcinoma.
infection to her own offspring. Thus, the virus needs some Before moving on, it would be worthwhile to under-
means for persisting in the face of virus-specific antibod- score several key differences between infections involving
ies and cytotoxic T cells of adaptive immunity. A means HCV and those infections involving HBV. In each in-
by which HCV is thought to evade adaptive immune de- stance, the immune response plays a major role in (i) con-
fenses is revealed by the following observations. First, trolling the infection, (ii) clinical recovery and protective
compared to the high fidelity of cellular DNA replication, immunity, and also (iii) pathogenesis. However, when
all viruses replicate their genomes with relatively low fi- adults are infected by HCV, the outcome is generally
delity. Moreover, the error rate is vastly greater in the case chronic infection, whereas infection of adults with HBV
of RNA viruses (i.e., one error in every 104 nucleotides in generally results in protective immunity. Thus, these two
the case of HIV, and about one error in every 105 nucle- viruses must interact differently with the immune system.
otides in the case of most other RNA viruses). Thus, if one Although much remains to be learned about the interac-
were to sample the HCV population in any infected indi- tions of these viruses with host immunity, we might note
vidual, one would find the simultaneous presence of a that HBV, which has a double-stranded DNA genome,
large number of genetic variants with different antigenici- evades immune defenses by generating an excess of empty
ties. Second, if one were to monitor an infected individu- decoy particles that soak up neutralizing antibodies. HCV,
als liver enzyme levels in the blood, which are indicative which has a single-stranded RNA genome, uses its inher-
of virus-mediated liver damage, one would notice cycles ently error-prone RNA polymerase to generate large num-
of approximately 6weeks in length. Together, these obser bers of antigenic variants (immune escape mutants). Yet
vations suggest that every 6weeks or so, the error-prone these are surely not the only viral immune escape and sur-
virus polymerase manages to generate immune escape vival strategies at play in each instance (Rehermann and
mutants that are able to evade the antigen-specific adap- Nascimbeni, 2005). Another difference to note is that he-
tive defenses that are presently expressed by the host. patocellular carcinoma in hepatitis C patients almost al-
These mutants prevail until the adaptive immune system ways develops against a background of cirrhosis, whereas
catches up with new antibody-producing B cells and cyto- in hepatitis B patients it might develop in the absence of
toxic killer T cells, at which time the cycle repeats. cirrhosis. Thus, there are different virus-specific factors
Importantly, patients infected with HCV as adults, who that contribute to hepatocarcinogenesis in each instance.
have become chronically infected, rarely recover. Yet when The differences noted above regarding the known
they do recover, it is not clear whether or not the infection means by which HBV and HCV evade host adaptive im-
is ever eliminated. (Recall that it also is not clear whether munity are relevant to the development of practical coun-
HBV-infected individuals ever clear that infection.) On termeasures against these viruses. For example, since HBV
the one hand, HCV reactivation has not been reported in produces essentially one serotype, it was possible to de-
those rare recovered individuals who subsequently un- velop an effective vaccine against that virus. In contrast,
dergo immunosuppression (compare to HBV infections, the vast number of HCV serotypes that are generated may
above). But on the other hand, by means of very sensi- preclude the development of a vaccine against that agent.
tive polymerase chain reaction (PCR) techniques, viral JC virus is a member of the polyomavirus family of
sequences have been detected in peripheral-blood lym- small, double-stranded DNA viruses (Chapter 15). (JC vi-
phocytes of recovered patients. It is not known whether rus is closely related to simian virus 40 [SV40], a virus that
or not such individuals might be infectious. (There is evi- has been extensively studied for its ability to induce neo-
dence that some patients who underwent antiviral drug plastic transformation of cells in culture and tumors in
therapy may have been cured of their HCV infections laboratory animals. See Chapter 15 for a detailed account
[Chapter 7].) As in the cases of patients who recover from of JC virus and SV40.) JC virus is ubiquitous in humans,
chronic HBV infections, patients who recover from HCV infecting more than 85% of adults, with most infections
infections likewise mount strong, adaptive antiviral im- occurring by the age of 15years. The virus is believed to
mune responses. initially infect the respiratory tract and to then dissemi-
Chronically infected HCV carriers usually live symptom nate to the kidneys and lymphoid organs. Moreover, JC
free for many decades, implying that the viral immune eva- virus establishes a lifelong chronic infection, during which
sion strategies also modulate immunopathology. However, it may replicate in tubular kidney cells. The primary and
in about 20% of chronic carriers, sufficient accumulated secondary infections, as well as the chronic infection, are

usually not associated with any illness. However, in indi- immunosuppression known. Indeed, because of the grow-
viduals who are severely immunosuppressed, JC virus can ing size of the AIDS population, PML has become a much
spread to the central nervous system, where it causes the more common condition in recent years, developing in as
usually fatal neurodegenerative disease progressive multi- many as 5% of all AIDS patients and presently affecting
focal leukoencephalopathy (PML). The term leukoen- nearly 1in every 200,000 persons overall. Once underway,
cephalopathy refers to the fact that lesions are seen only it is relentlessly progressive, with death generally occur-
in the white matter. ring 3 to 6months after the onset of symptoms.
Historically, PML is the first demyelinating disease in It is not known with certainty how JC virus invades the
humans that was found to be caused by a conventional central nervous system in order to give rise to PML, al-
virus (i.e., in contrast to the transmissible spongiform en- though some researchers suggest that it may do so via
cephalopathies, which are caused by proteinaceous infec- infected B cells. Once in the central nervous system, JC
tious particles, or prions; see below). PML was an extremely virus has a tropism for oligodendrocytes (the myelin-
rare condition when it was first identified as a distinct dis- generating cells in the brain) and astrocytes (star-shaped
ease entity in the 1950s. At that time, it was essentially cells that support and nourish nerve cells in the brain)
seen only in organ transplant recipients on immunosup- (Fig. 3.10). Lytic JC virus infection of oligodendrocytes
pressive regimens and in cancer patients whose neo- causes the demyelination characteristic of PML lesions.
plasms compromised immunity (e.g., Hodgkins disease The receptor for JC virus was recently identified as
and chronic lymphocytic leukemia). Immunocompetent 5-HT2A, a receptor of the family of serotonin receptors.
individuals who are chronically infected with JC virus gen- The serotonins are neurotransmitters, and consequently,
erally do not develop PML, presumably because their JC this finding neatly accounts for the neurotropism of JC
virus-specific CD4 and CD8 T cells control the infection. virus. (Incidentally, 5-HT2A also is the receptor for the
These virus-specific T cells likewise play a role in deter- hallucinogen LSD.)
mining the clinical course of PML, since disease progres- How might we classify JC virus infections? On the one
sion is associated with declining T-cell levels. hand, the generally benign persistent JC virus infections
At present, the majority of PML cases occur in pa- in immunocompetent individuals might be classified
tients with AIDS, which is the most severe condition of as chronic in nature. On the other hand, the emergence

Figure3.10 The cellular organization of the central nervous system. Modified from M. McKinley and V. OLoughlin,
Human Anatomy (McGraw-Hill, New York, NY, 2006), with permission.

White matter
Grey matter




Myelinated axon

Ependymal cells
Myelin sheath

Ventricle of
the brain
Modes of Virus Infection and Disease 73

of PML in immunocompromised individuals might be to the surface of the epithelium and differentiate, produc-
thought of as a slow virus disease, as well as a chronic in- tive virus replication occurs, thereby facilitating the effi-
fection. In any case, JC virus infections cannot be neatly cient transmission of the infection to new individuals. All
categorized as either chronic or slow. the while, viral genomes are maintained in the basal layer,
Papillomaviruses are small double-stranded DNA vi- thereby sustaining the long-term persistent infection. In-
ruses that cause benign skin warts, and more importantly, terestingly, replication of papillomavirus genomes in the
they are the primary cause of carcinoma of the cervix, one basal cells is regulated by the fact that the viral origin of
of the most common cancers in women (Chapter 16). DNA replication (ori) sequence interacts like a cellular ori
These pathologies may start to develop several months to with the cellular DNA replication machinery. In this way,
years after the initial infection. Indeed, invasive cervical the average number of papillomavirus genomes per basal
carcinomas may emerge as late as 20 to 50years after the cell remains constant over time.
initial papillomavirus infection. Like other viruses that establish persistent infections,
The establishment and maintenance of chronic papil- papillomaviruses likewise have evolved strategies that
lomavirus infections are part and parcel of the papilloma- enable them to evade immune clearance. For example,
virus replication strategy. In brief, these viruses have a these double-stranded DNA viruses transcribe only one
rather strict tropism for the basal cells that constitute the of their DNA strands. Thus, they do not inadvertently
lowest of the several layers of the epidermis (Fig. 3.1). generate the double-stranded RNA that might induce IFN
Papillomaviruses access these cells through tears in the (Chapter 4). Moreover, papillomaviruses do not harm the
skin and mucosal layers. Basal cells continually proliferate abortively infected basal cells, which sustain the persistent
in order to replenish the nonproliferating cells of the infection. For these two reasons, papillomaviruses do not
overlying layers, which continually die and slough off. As initially activate the innate immune system. And if the in-
basal cells divide, one daughter cell migrates into the up- nate immune system is not activated, there is no adaptive
per layers and begins a course of differentiation to a kera- response. (This key feature of adaptive immunity is ex-
tinocyte while the other daughter cell remains behind as a plained at length in Chapter 4.) Eventually, the adaptive
basal cell. Importantly, papillomaviruses do not carry out immune system is activated, presumably because some
a complete replication cycle in the basal cells. Instead, they virus-producing cells are killed. Nevertheless, the papil-
only replicate their genomes, producing only a few copies lomavirus stratagem of restricting late protein synthesis
in each basal cell. Then, these viral genomes are distributed and virus production to the terminally differentiated kera-
at random between the daughter cells when the basal cell tinocytes at the outermost layers of the epithelium places
divides. Again, one of the daughter cells remains behind these virus-producing cells beyond the reach of immune
in the basal layer, while the other migrates to the upper attack by cytotoxic T lymphocytes (CTLs), while also pre-
layers of the epithelium. venting viruses from being exposed to neutralizing anti-
As the basal cells migrate to the surface, they eventually bodies. (CTLs and neutralizing antibodies are the key
differentiate into keratinocytes, which acquire the ability to effectors of adaptive immunity [Chapter 4].) Overall, this
support the complete papillomavirus replication cycle. For strategy is so successful that at least 25million Americans
its part, the virus enhances the efficiency of its replication currently have active genital papillomavirus infections.
by producing two proteins, E6 and E7, which keep the kera- We conclude this section with a brief consideration of
tinocytes in a vegetative state; this is a necessary condition HIV, mainly to note that infections with this virus display
for virus replication, since the virus is dependent on the defining features of each of our three categories of persis-
cellular DNA replication machinery, which otherwise is ac- tent virus infections (i.e., slow, chronic, and latent). We
tive only in replicating cells. The mechanisms of action of begin by noting that the average time that elapses between
the papillomavirus E6 and E7 proteins are described in HIV infection and the onset of full-blown AIDS is about
Chapter 16. The mechanism by which certain papillomavi- 10years (Chapter 21), thereby qualifying AIDS as a slow
ruses induce cervical carcinoma is also described in Chap- virus disease. Next, infectious virus always is present and
ter 16. As you may have surmised, expression of E6 and E7 being shed, a characteristic of a chronic infection. Finally,
is required for the development of this cancer, but as will be at any given instant, there are many more latently infected
seen, there is more to this part of the story (Chapter 16). cells than productively infected cells, a characteristic of a
We can now appreciate that papillomavirus persistence latent infection (see below). Still, for me at least, HIV in-
results from a strategy in which these viruses infect the fections are most appropriately categorized as chronic.
basal cells beneath the skin and mucosa, in which limited The reasons are as follows.
viral genomic replication occurs and in which the infec- Latency in the case of HIV is distinctly different from
tion is otherwise abortive. But as these same cells migrate the quintessential latency that characterizes infections

with the human herpesviruses, such as HSV-1, a virus that those with oral HSV-1 infections and is shed from the
establishes lifelong latent infections, which periodically genitalia of those with genital infections.
may reactivate to give rise to recurrent acute episodes. In The acute phase of an HSV-1infection is generally lo-
contrast, while HIV is latent in most infected cells of an calized to the initial site of infection. That is so because
HIV-infected individual, a subset of infected cells is pro- the virus receptor, heparin sulfate, is also a major constit-
ducing virus at any given time, such that untreated indi- uent of the extracellular matrix. Thus, despite the fact that
viduals are always viremic. Moreover, herpesviruses spend heparin sulfate is present on nearly all cells, extracellular
most of their time hiding from the host immune system, virus cannot diffuse very far before being bound by hepa-
and at no time are they actually attacking it. In contrast, rin sulfate molecules of the extracellular matrix.
although HIV is hiding from the immune system in la- HSV-1has evolved numerous strategies for evading in-
tently infected cells, a striking feature of an HIV infection nate and adaptive immune responses during the acute
is the unrelenting direct attack by the virus on the im- phase of an infection (recounted in detail in Chapters 4
mune system. Additionally, the continued production and and 18). Although these viral countermeasures against
shedding of HIV and the delayed onset of AIDS, about host immune defenses are clever, perhaps even diabolical,
10 years after the initial infection on average, are more within about 2weeks time the immune system generally
typical of chronic infections, hence the inclusion here of does succeed in resolving the acute phase of the infection.
HIV under chronic infections. Nevertheless, and importantly, the viral immune evasion
Some of the numerous mechanisms used by HIV to strategies provide a window of opportunity for the virus to
evade innate and adaptive immune responses are recounted establish latent infection. This occurs as follows. During the
in Chapters 4 and 21. HIV persistence and the pathogen- acute phase of the HSV-1 infection, the virus reproduces
esis of AIDS are intimately related to the molecular biology quickly in the epithelial cells at the initial site of infection.
of the virus and its interaction with its host cells. These Some of the progeny virus that is produced there then in-
issues are considered in detail in Chapter 21. vades the endings of the sensory neurons that innervate the
epithelium. These virus particles are conveyed by retro-
Latent Infections grade axonal transport to cranial or spinal sensory ganglia,
As mentioned above, herpesviruses give rise to quintes- where the viral genomes persist essentially indefinitely in
sential latent infections. The known human herpesvi- the nuclei of a minority of neurons (Fig. 3.4 and 3.11).
ruses are HSV-1 and -2, human cytomegalovirus, EBV, The establishment of HSV-1 latency would be note-
VZV, human herpesvirus 6 (HHV-6), HHV-7, and the worthy if it merely provided some advantage to the virus
more recently discovered HHV-8, which is associated or if it were merely clinically relevant. But it is both advan-
with Kaposis sarcoma in AIDS patients (Chapters 18 tageous to the virus and clinically relevant. We begin by
and 21). considering the advantage to the virus of its latency life-
Each of the human herpesviruses has evolved a some- style. Bear in mind the following points: (i) HSV-1 is
what different strategy for establishing latency, in part be- transmitted by intimate contact between individuals,
cause each targets a particular host cell type in which to (ii) the acute phase of the infection is resolved within a
do so. HSV-1, HSV-2, and VZV establish latency in neu- time interval of 2weeks, (iii) humans are generally selec-
rons, EBV in B lymphocytes, cytomegalovirus in mono- tive regarding their intimate contacts, and (iv) humans
cytes and lymphocytes, and HHV-6 and HHV-7in T lym- are the only host for HSV-1. Considering these points,
phocytes. The cell type in which HHV-8 becomes latent is HSV-1 would be in danger of not surviving in Nature un-
not yet known. Regardless, the herpesvirus strategy for es- less it had some special strategy for perpetuating itself in
tablishing and maintaining latency depends on cellular its host population. This special strategy is the ability of
factors, such as whether the host cell might be dividing the virus to establish latency in the ganglia of neurons that
(which it is not, in the case of neurons) or long-lived innervate the initial site of infection. These neurons pro-
(which it is not, in the case of lymphocytes). Infections by vide an immunologically privileged site, in which the virus
each of the herpesviruses are considered in some detail in might find a long-term safe haven. The immunologically
Chapter 21. Our focus here is on HSV-1, as an example of privileged status of neurons is due in part to the fact that
a classic latent infection. they express only low levels of major histocompatibility
HSV-1, which gives rise to both oral and genital herpes, complex class I molecules (which display virus antigens to
infects more than 60% of all Americans by late adoles- cytotoxic T cells [Chapter 4]). Moreover, the virus con-
cence. The virus replicates in epithelia and is usually tributes to its stealthiness in neuronal cells by generating
transmitted by direct intimate contact between individu- only negligible amounts of proteins that might be recog-
als, reflecting the facts that it is present in the saliva of nized by immune effectors (Box3.8).
Modes of Virus Infection and Disease 75

Primary infection
Herpes simplex virus type 1 Varicella virus
Latent virus Neuron in dorsal
root ganglion

Spinal cord

Viral transit up
peripheral nerve

Mild pharyngitis Chickenpox

or stomatitis

Fever Age
Sunlight to face X-irradiation Activation
Menstruation (act via depressed CMI) of virus in
Nerve section at * neuron

Spinal cord

Viral transit up
peripheral nerve

Cold sore Zoster (shingles)

Figure3.11 Mechanisms of latency in herpes simplex and VZV infections. Primary infection occurs in childhood or
adolescence; latent virus is located in the cerebral or spinal ganglia. Reactivation of HSV causes recurrent herpes
simplex; reactivation of VZV causes zoster (shingles). CMI, cell-mediated immunity. Adapted from D. O. White and
F. J. Fenner, Medical Virology, 4th ed. (Academic Press, Inc., San Diego, CA, 1994), with permission.

Now here is an important point that is relevant both to the original site of infection, where it can undergo a new
the natural history of HSV-1 and to herpesvirus disease. It cycle of lytic replication in the epithelium and thus be
would do HSV-1 little good if the latent phase of the in- transmitted to new individuals (Box3.10).
fection were permanent, since latent virus cannot be A key feature of the above HSV-1 strategy is that it
transmitted to susceptible individuals. Consequently, la- makes use of two cell types: epithelial cells, in which the
tent HSV-1 genomes are able to periodically reactivate, virus undergoes productive infection, and neuronal cells,
thereby reinitiating the productive infection (Fig. 3.11). It in which it establishes latency. For years it was believed
is not yet entirely clear how reactivation occurs at the mo- that during the latent periods no infectious virus was pro-
lecular level (Box3.9). However, reactivation is known to duced but viral genomes instead remained cryptic in cer-
be triggered by hormonal factors (e.g., the menstrual cy- tain protected cells. However, newer, more sensitive mo-
cle), as well as by external factors such as psychological lecular procedures show that infectious HSV-1 may be
stress, physical trauma, and ultraviolet (UV) irradiation shed continuously by asymptomatic individuals. Thus, la-
(i.e., time spent in bright sunlight). In addition, reactiva- tency in the case of HSV-1 provides a means for an in-
tion is likely to occur when an individual is immunosup- fected individual to be a continuing source of infection
pressed. At any rate, when HSV-1 reactivation occurs, the for a lifetime, perhaps accounting for the ability of this
virus replicates in some of the neurons in which it has virus to be so ubiquitous in the human population. Two
been latent. It then is transported back down the axon to other points to note are as follows. First, transmission by

Box 3.8 Box 3.10

Whether the latent viral genome might persist as an It is worth reiterating that any virus that gives rise only to
episome or is integrated by covalent bonds into the cel- acute infections faces the dilemma that eventually all the
lular chromatin is one of the defining characteristics of a individuals in its host population may have been infected
particular latent infection. In the case of HSV-1, the viral and thus become immune. For this reason, acute viruses
genomes persist as free episomes in neuronal cell nuclei, require some minimal host population size to allow for
approximately 10 to 100 genomes per nucleus. But strictly a sufficient number of new susceptible individuals to be
speaking, by definition an episome can exist either born into it to sustain the virus. In the case of measles
autonomously or as part of a chromosome. In contrast, virus, that minimal population size is about 500,000indi-
a plasmid is an extrachromosomal DNA molecule that is viduals. Thus, measles virus, which has no animal reser-
capable of replicating independently of the chromosomal voir, can only recently (i.e., within the last 10,000years or
DNA. Although it may be more appropriate to refer to so) have become a human virus, since such large human
the latent extrachromosomal HSV-1 genome as a plas- populations did not exist before then (Chapter 11). On
mid, the journal literature often refers to it as an episome. the other hand, latency provides a means for a virus to
sustain itself in populations of much smaller size. Do you
It is enlightening to compare the state of latent HSV-1 see why that is so? HSV-1has long been a human virus,
genomes in neuronal cells with the state of latent HIV coevolving with its host before and ever since humans
genomes in CD4helper T cells. Whereas latent HSV-1 diverged to become a distinct species.
genomes persist in neuronal cells as extrachromosomal
episomes, latent HIV genomes persist in CD4helper
T cells in an integrated state. Why the difference? Per-
haps the reason is as follows. Neurons do not replicate, asymptomatic individuals is more important than you
whereas CD4helper T cells do. By integrating its genome
might think at first, since many candidates for infection
into the cellular chromatin, HIV ensures that it will be
might resist some kinds of intimate contacts with those
passed on to each daughter cell at division. HSV-1has no
expressing obvious cold sores or other herpetic lesions,
such concern in nondividing neurons. Alternatively, or in
but not with those who are asymptomatic. Second, these
addition, the integration of the HIV genome is a crucial
step that enables it to subsequently be reactivated and
newer findings also somewhat blur the distinction be-
replicated (Chapter 21). tween chronic and latent infections.
Finally, coming full circle back to portals of entry and
transmission, with which this chapter began, the ability of
HSV-1 to thrive in the human population demonstrates
how strongly we crave intimate contact. We could not avoid
Box 3.9 becoming infected by acute respiratory viruses, such as
the rhinoviruses and influenza virus, short of becoming
The expression of HSV-1 genomes is not completely shut hermits. But we could avoid becoming infected by HSV-1
down in latently infected neurons. Indeed, the establish- simply by avoiding intimate physical contacts. It is not a
ment and maintenance of latency require the expres- stretch to presume that HSV-1 evolved to exploit our
sion of specific HSV-1 genes in those cells. The family
human urges for intimate contacts.
of HSV-1messenger RNAs (mRNAs), referred to as the
latency-associated transcripts, is specifically expressed
Transmissible Spongiform Encephalopathies:
in latently infected neurons. Moreover, latency depends
on cellular factors as well. Actually, it may depend on the
The transmissible spongiform encephalopathies, also re-
absence of cellular factors, such as particular cellular tran-
ferred to as subacute spongiform encephalopathies, are an
scription factors that might be needed for the expression
of crucial viral early genes. With appropriate stress to
especially intriguing group of slow, infectious, neuro
these cells, the necessary cellular genes might be activated degenerative diseases. Scrapie in sheep is the prototype for
so that viral reactivation might occur in at least a few this group of diseases, which also includes kuru, Creutzfeldt-
neurons. Also, reactivation from latency may require the Jakob disease (CJD), fatal familial insomnia, and the
expression of specific reactivation-associated viral genes Gerstmann-Strussler-Scheinker syndrome in humans
as well (Fig. 3.11). Current understanding of the work- and bovine spongiform encephalopathy (BSE) in cattle,
ings of these viral factors is recounted in Chapter 18. known colloquially as mad cow disease. (Actually, fatal fa-
milial insomnia and the Gerstmann-Strussler-Scheinker
Modes of Virus Infection and Disease 77

syndrome are inherited, rather than infectious, diseases; triggered by mutation(s) in the prion protein gene. For ex-
see below.) ample, in the Gerstmann-Strussler-Scheinker syndrome, a
Importantly, although the agents responsible for these change in PrPC amino acid 102 from proline to leucine is
diseases easily pass through filters, they are not viruses! found in most affected individuals. Sporadic cases of these
What then are they? First, let us briefly consider what we diseases are then caused by random, spontaneous events.
mean by the term virus. Admittedly, we do not have any What sort of experimental evidence might support the
consensus for an all-inclusive definition for a virus (Chap- above prion model? In one early study, transgenic mice
ter 1). Yet we likely would agree that viruses contain genes, were generated which lacked both copies of the mouse
have an apparent structure, and elicit an immune re- PrPC gene. These mice were completely resistant to mouse
sponse. Unlike the agents that we agree are viruses, the scrapie prions. However, when hamster PrPC genes were
agents responsible for the transmissible spongiform en- introduced into these mice so that they expressed hamster
cephalopathies have no apparent genome or virus struc- PrPC, these mice then were susceptible to hamster scrapie
ture, nor do they elicit an immune response. Moreover, prions but remained resistant to mouse scrapie prions.
they are resistant to physical and chemical treatments (e.g., Presumably, hamster scrapie prions, but not mouse scrapie
heat, radiation, and formaldehyde) that would inactivate prions, could interact with the hamster PrPC to catalyze
a conventional virus. These facts, while intriguing, do not their conversion to the scrapie isoform.
tell us the nature of the agents responsible for the trans- More recently, Prusiners research group produced
missible spongiform encephalopathies. So then, what are neurologic dysfunction in mice by inoculating them in-
they? tracerebrally with a recombinant mouse prion protein
The agents responsible for the transmissible spongi- generated in Escherichia coli (Legname etal., 2004). The
form encephalopathies actually appear to be variant forms recombinant mouse prion protein contained a mutation
of normal cellular proteins. Both the normal and infec- analogous to the human prion protein mutation that causes
tious forms of these proteins are referred to as prions, an Gerstmann-Strussler-Scheinker syndrome (see above).
acronym derived from proteinaceous infectious particle. The preclinical phase (incubation period) for scrapie
This name was coined by Stanley Prusiner, who won a in sheep is about 3years. For CJD and kuru in humans, it
Nobel Prize in 1997 for his breakthrough studies into the may be as long as 30 years. But once symptoms appear,
proteinaceous nature of prions. The aberrant prion confor- patients generally succumb within a year.
mation responsible for scrapie is termed PrPSc. (Although Clinically, the subacute spongiform encephalopathies
PrPSc now designates scrapie prion protein, the origin are characterized by a loss of muscle control, shivering,
of the PrP abbreviation is protease resistant protein.) tremors, and loss of coordination, and in humans demen-
The corresponding normal host cell prion protein is tia is also one of the characteristic symptoms. The term
termed PrPC. Both PrPSc and PrPC are about 27,000 to spongiform refers to the characteristic vacuolated neu-
30,000 daltons (Da). However, differences in folding rons. Histological changes also include the proliferation
between PrPSc and PrPC account for the differences in and hypertrophy of astrocytes and the fusion of neurons
their biological effects. PrPC is predominantly -helical, with adjacent glial cells. PrPSc is taken up by neurons and
whereas PrPSc is a -sheet-rich structure. probably contributes to the vacuolization of the brain tis-
Prusiners studies demonstrated to the satisfaction of sue. Interestingly, prions are found in other tissues as well,
many researchers (not all are convinced) that PrPSc repli- but only the brain shows any signs of disease. It is not
cates and causes scrapie by somehow catalyzing the con- known how prions cause the fatal pathogenesis of the
version of the benign cellular PrPC conformation to the spongiform encephalopathies.
PrPSc conformation. Once underway, this conversion pro- Although kuru is a disease that readers of this book will
cess might escalate exponentially. The accumulation of probably never actually see, it makes for a most compel-
the PrPSc isoform in the brain then leads to the disease. ling story. This transmissible spongiform encephalopathy
The fact that the aggregates of PrPSc that accumulate are is limited to a tribe of Fore-speaking people, who live in the
comprised of host proteins may account for why there is remote mountains of eastern New Guinea. The term kuru
no immune response to these agents in infected animals means shivering or trembling in the Fore language.
and patients. In the late 1950s, as the Fore people were just emerging
In the transmissible spongiform encephalopathies, from the Stone Age, stories about their strange disease be-
such as kuru and CJD, the transformation of PrPC to PrPSc gan to leak out to the modern world. These stories in-
is initiated by an exogenous source of PrPSc. In the inherited trigued virologist Carlton Gajdusek at the U.S. National
forms, such as fatal familial insomnia and the Gerstmann- Institutes of Health, who came to the Fore people to see
Strussler-Scheinker syndrome, the transformation is the disease for himself. What he found was as follows. The

Fore people numbered about 35,000individuals, and kuru the Fore people for several years. This cycle finally was
affected about 1% of their population. Importantly, the broken when cannibalism among the Fore people ended,
disease occurred primarily in women, with many fewer in the 1960s.
cases seen in children (bear in mind the generally long Gajduseks discovery that kuru is a transmissible dis-
incubation periods for these diseases, which explains the ease took on new importance when the more widespread
rarity of disease in children; see below). Interestingly, chil- human presenile dementia, CJD, was found to be caused
dren of both sexes were affected in equal numbers. How- by an agent strikingly similar to the agents responsible for
ever, the disease was most prevalent in women, and for a scrapie and kuru. CJD has a frequency of about one per
time the majority of all deaths in women of the tribe were million individuals worldwide. Interestingly, CJD cases
due to kuru. Consequently, men outnumbered women by have been traced to medical interventions such as neuro-
3 to 1in some villages. surgery, corneal transplants (a donor was diagnosed post-
Gajdusek was puzzled by (i) the strange demographics mortem with CJD), and administration of growth hor-
of kuru in the Fore population, (ii) its slow progression, mone derived from pituitary extracts.
and (iii) the fact that there was no inflammation or fever In addition to the spread of CJD by a transmissible
associated with the condition. For these reasons, he thought agent, the disease also may appear suddenly, without any
that the disease might be caused by genetic factors, or obvious epidemiological signs. Those random instances
perhaps by malnutrition. Then, while Gajdusek was still are referred to as sporadic CJD. Presumably, sporadic CJD
searching for the cause of kuru, William Hadlow, an occurs when a PrPC isoform spontaneously assumes the
American who had been studying scrapie, entered the PrPSc conformation. Perhaps kuru originally arose among
scene. Fortuitously, while Hadlow was visiting the Well- the Fore people from the consumption of an individual
come Medical Museum in London, he happened to look with sporadic CJD.
at a display on kuru that Gajdusek had prepared. It was Fatal familial insomnia, which is a very rare inherited
there that Hadlow noticed that the pathological changes prion disease due to a specific mutation in the PrPC gene,
occurring in human kuru were virtually identical to those differs from other prion diseases in that it predominantly
seen in ovine scrapie. Importantly, it was already well affects one area of the brain, the thalamus, which influences
known that scrapie is due to a transmissible agent. sleep. Affected individuals eventually become completely
Following up on Hadlows insight, Gajdusek attempted sleepless. Like other prion diseases, this condition is invari-
to transmit kuru to chimpanzees. His success in that effort ably fatal. Note that the dominant gene responsible for this
demonstrated that kuru too is caused by a transmissible disease has been found in only 28 families worldwide.
agent. Moreover, it was the first demonstration of a human BSE (referred to informally as mad cow disease) made
slow infection, conventional or otherwise. The incubation its first appearance in 1987, in Great Britain. Interestingly,
period for kuru in the chimpanzees ranged from 14 to a cannibalism of sorts may underlie Britains mad cow
82months, and the disease could be serially propagated outbreaks; cannibalism, that is, between cows and sheep.
from one chimpanzee to another. (Do you think that the Scrapie in sheep, which is very similar to BSE in cattle, has
study of kuru transmission might have been a desirable existed in England for at least 200years, where it affects
topic for a Ph.D. dissertation?) 0.5 to 1% of the sheep population per annum. Here now
Considering that kuru is caused by a transmissible is the connection between cannibalism and BSE. For
agent, rather than an inherited mutation, what might ex- several decades, cattle feed had included protein supple-
plain its strange demographics in the population? The an- ments made from the carcasses of other animals, includ-
swer in a word is cannibalism. The Fore people were can- ing sheep (this material is referred to as offal). For most
nibals, and kuru was transmitted from person to person of that time, scrapie never crossed into cows. However,
by cannibalism. Actually, the Fore people ate their dead this changed in the 1980s because of a new way of render-
relatives as an act of homage during funeral rites. The ing the carcasses, which led to feed supplements that
bodies of the deceased would be cut up into parts. As were contaminated with the scrapie agent. In earlier times,
might be expected, the men took the meatiest parts for carcasses were exposed to prolonged heat treatment,
themselves, leaving the pancreas, liver, kidneys, and, im- which apparently was sufficient to inactivate PrPSc to safe
portantly, the brain for the women and the children. Many levels. In the 1980s, there was a changeover to rendering
of the women would develop kuru after eating the in- the carcasses by a chemical procedure. This was done in
fected brains. When they died, they too were ritually eaten, order to cut costs; the heat treatment required the burn-
thereby escalating the epidemic. ing of increasingly expensive fossil fuels. Importantly, the
Gajdusek came to realize the connection between new procedures apparently were not sufficient to inacti-
kuru and this ritualistic cannibalism after living among vate PrPSc.
Modes of Virus Infection and Disease 79

BSE has been in the news in recent years, not only for accusations that the interests of the British meat industry
its effect on cattle but also because of evidence that eating were placed above those of the public. (Indeed, one inves-
meat from infected cows can give rise to CJD in humans. tigating British scientist is quoted as saying, And if we had
Because CJD is so rare (even among those exposed to been too alarmist, we were in danger of upsetting the whole
BSE-contaminated beef), it is difficult to demonstrate a of the meat industry in Britain and elsewhere in Europe.)
statistically significant increase in the number of CJD Yet outside England, restrictive steps were taken. European
cases that might be due to eating contaminated meat. countries and the United States imposed a ban on the im-
However, the change in the demographics of CJD in Great port of British cattle. Finally, in response to public pressure,
Britain is striking, and it strongly implies that something the British government banned the beef-processing indus-
new is amiss. Previously, CJD was a disease of the elderly. try from the practice of feeding cows and sheep back to
But with the emergence of the BSE epidemic in British each other. In addition, cattle already showing signs of BSE
cattle, CJD in Great Britain emerged in young adults. were excluded from use in human food and destroyed.
Moreover, the PrPSc isoform found in the younger CJD (What is your reaction to the fact that belated governmen-
patients has a glycosylation pattern similar to that found tal intervention was necessary to impose these changes?)
in cattle with BSE, and it is significantly different from the
isoform seen in older CJD patients. Selected Readings
Because (i) BSE has a long incubation period in cattle
Igakura, T., J. C. Stinchcombe, P. K. C. Goon, G. P. Taylor, J. N.
and (ii) CJD has a long incubation period in humans, the Weber, G. M. Griffiths, Y. Tanaka, M. Osame, and C. R. M.
problem of BSE in British cattle and its danger to humans Bangham. 2003. Spread of HTLV-I between lymphocytes by
became apparent only recently. Alarmingly, however, for virus-induced polarization of the cytoskeleton. Science 299:
2years after BSE became known, infected cattle were still 17131716.
allowed into Englands food supply. Legname, G., I. V. Baskakov, H. O. Nguyen, D. Riesner, F. E. Cohen,
S. J. DeArmond, and S. B. Prusiner. 2004. Synthetic mammalian
Some of the early complacency surrounding BSE may prions. Science 305:673676.
have been due to the fact that whereas scrapie had been Rehermann, B., and M. Nascimbeni. 2005. Immunology of hep-
around in sheep for several centuries, it was not known to atitis B virus and hepatitis C virus infection. Nat. Rev. Immunol.
cause a problem in humans. However, there have been 5:215229.

Host Defenses and Viral

Cytokines: the IFNs
Cytokines: TNF-a, Some Other Cytokines,
and Inflammation
Macrophages, Neutrophils, NK Cells, and
Antibody-Dependent Cellular Cytotoxicity
The Complement System
Viral Evasion of Innate Immunity
Evasion of IFNs
Evasion of Cytokines
Evasion of NK Cells and ADCC
Evasion of Complement
APOBEC3G and the HIV Vif Protein
Antibodies and B Cells
Antibody diversity Much of modern virology concerns the molecular biology of viral gene
Viral Evasion of Antibodies expression and virus replication. Indeed, that is the emphasis in this book.
Cell-Mediated Immunity However, it is important to bear in mind that the environment for a mam-
Antigen Presentation by MHC malian virus consists of more than just the susceptible cells of the host in
Class I Molecules
which the particular virus replicates. Indeed, other specialized cells and
Antigen Presentation by MHC
Class II Molecules tissues of the host are responding to the infection to contain and eliminate
The Rationale for MHC Restriction it. Moreover, even the susceptible cells themselves respond to control the
Activation of Th Cells: Dendritic Cells and infection. Therefore, we need to take account of the entirety of the defen-
B Cells sive stratagems of the host and of the countermeasures that viruses have
Activation of B Cells evolved in order to evade and even subvert them.
Activation of CTLs
We begin by reminding ourselves that from the point of view of their ver-
Mechanism of Action of CTLs
T Cells and Antiviral Cytokines
tebrate hosts, viruses, as well as other pathogenic microbes, are potentially
Viral Evasion of Cell-Mediated Immunity lethal invaders. Consequently, over the course of hundreds of millions of
Inhibition of antigen presentation to CTLs years, vertebrates have had to evolve defenses not only against viruses but
Inhibition of antigen presentation to helper also against bacteria, fungi, protozoa, and metazoan parasites (e.g., worms).
T cells Each of these different classes of invasive pathogens uses a different
Inhibition of apoptosis strategy to establish itself in the host. Moreover, there are a variety of strat-
IMMUNOLOGICAL MEMORY egies used by different pathogens within each class, as well as between
SELF TOLERANCE classes. In response, vertebrates have evolved a network of multiple defen-
THE IMMUNE SYSTEM IN DISEASE sive measures, and they are able to bring to bear the most appropriate
combinations of these defensive measures to counter each particular class
Autoimmune Disease
of invader. Although we are primarily concerned with host defenses
against viruses, note that some but not all antiviral defenses overlap de-
fenses used against other classes of invaders as well.
Our antivirus defenses indeed are formidable. This may reflect the fact
that viruses can evolve much more quickly than their more slowly evolv-
ing vertebrate hosts. Thus, the host needs to have potent and comprehen-
sive defenses in place that might cope with whatever countermeasures the
rapidly evolving viruses might adapt.
One of my hopes in writing this chapter has been to convey just how
formidable, and one might say refined and elegant, our immune defenses
Host Defenses and Viral Countermeasures 81

are. Yet, as we just have noted, rapidly evolving viruses Key differences between the innate and adaptive im-
continue to devise countermeasures that challenge those mune systems are as follows. The innate immune system
host defenses. Thus, much as one may be struck by the recognizes general features present on a variety of patho-
potency and complexities of the host defenses, many of gens, and it does so by means of a fixed collection of un-
the most intriguing viral adaptations are those that enable changing receptors which are encoded in the germ line of
these microbes to evade and even subvert those host de- the host. These receptors recognize specific ligands that
fenses. Indeed, only those viruses which have succeeded in are present on a variety of different pathogens (e.g., bacte-
doing so can be sustained within any host population. rial lipopolysaccharides). In contrast, the adaptive im-
Thus, I hope that after reading this chapter you will have a mune system uses receptors that are specific for each par-
high regard for just how formidable our antiviral defenses ticular pathogen that might be encountered. The receptors
actually are, and in turn, I hope you also will appreciate of the adaptive immune system are not encoded in the
the variety of countermeasures that viruses have evolved germ line and must be generated by a special process, de-
to evade those defenses. The variety of these countermea- scribed below. Because of the difference in the origins of
sures is vast. For example, among the herpesviruses (a fam- the receptors of the innate and the adaptive immune sys-
ily of large DNA viruses), perhaps one-half of the entire tems, the innate immune system employs a repertoire of
complement of the 80 or more viral genes are concerned relatively few receptors, each of which is always present in
primarily with immune evasion. (If a virus were able to plentiful amounts, whereas the adaptive system has a very
completely countermand our immune defenses, then nei- large repertoire of receptors, but the number of receptors
ther we nor the virus would still be here. Importantly, the with any particular specificity is initially low.
virus needs to evade immune clearance only long enough Each cell of the adaptive immune system expresses im-
to pass the infection on to a new susceptible individual.) mune receptors of only a single specificity. However, cells
It is impossible to cover all of the important details of of that specificity will undergo a dramatic expansion if
the mammalian immune responses and all of the viral im- they should encounter an invader that is recognized by
mune evasion strategies in just one chapter, albeit a long their particular receptor. Indeed, the adaptive immune
one. Instead, we consider a few examples of some of the system is so named because its responses are tailored to
best-understood or especially intriguing (at least to the each particular pathogen. In fact, the specificity and ef-
author) viral evasion strategies. Some of these, as well as fectiveness of the adaptive immune response are continu-
ones not covered here, are discussed in more detail in the ously refined throughout the course of the infection.
chapters that follow. (Many readers will likely be more The differences between the innate and adaptive im-
than satisfied with the amount of detail crammed into mune systems outlined above have the following conse-
this chapter. However, rest assured that it is but a sum- quences. First, the innate immune system can begin to
mary of how the immune system works, hopefully con- effectively control an infection within the first several
taining enough details to impress you with its elegance hours. In contrast, the adaptive immune system has to be
and potency, and with the sophistication of the many viral activated by a special process to expand the number of
countermeasures against it.) effectors that can respond to the particular pathogen be-
ing encountered. For this reason, the adaptive immune
system is not effective until the fourth day after infection
OVERVIEW OF DEFENSES or later. Another important feature of the adaptive re-
The first aim of our antiviral defenses is to prevent infec- sponse is that it has a memory of past encounters, which
tion. If infection should occur, the purposes are then to leaves it primed to react quickly should it reencounter a
slow the replication and spread of the virus, followed by pathogen that it responded to earlier. This memory is the
its elimination. Towards those ends, our defenses may be basis for immunity. Indeed, one objective of vaccination is
thought of as comprised of a three-tiered system. The first to generate memory. The innate immune system has no
of these tiers consists of the physical barriers presented by memory of past infections.
the skin and by the mucosae that line the respiratory, the Despite the distinctions between the innate and adap-
gastrointestinal, and the urogenital tracts. The skin and tive systems, separating them is somewhat artificial, since
the aforementioned mucosae provide an effective shield they frequently work together. Indeed, a major difficulty
that prevents most viruses from gaining access to the cells in describing the immune system is that it is a network
and tissues of the body. Should a virus invade or penetrate that involves interactions between many different players.
one of these physical barriers, it then confronts the second Thus, in describing the individual players, it is important
tier of defense, the innate immune system, and if neces- not to lose sight of the forest for the trees. To merely cata-
sary, the third tier, the adaptive immune system. log the actions of each individual player would miss the

important interactions that together constitute the im- replicating, these properties of the vaginal epithelium
mune systems response to the infection. Because of these combine to create an effective barrier against most viruses.
interactions, we will see individual players reemerge mul- Moreover, the vagina is well colonized in healthy subjects
tiple times, as we introduce new players with which they by a bacterial flora that produces lactic acid. The resulting
interact. acidity, as in the case of the gut, protects the vagina from
infection by most viruses. In addition, the epithelial cells
that line the reproductive tract, like those that line the re-
PHYSICAL BARRIERS AGAINST spiratory and gastrointestinal tracts, are covered by a layer
INFECTION of protective mucus. Nevertheless, the vaginal epithelium
The skin is an effective barrier against viruses in general. can be torn, for example, during sexual intercourse, pro-
That is so because viruses require living cells to support viding a site of infection for viruses such as genital pa
their replication, and the skin is covered with multiple lay- pillomaviruses, herpesviruses, and human immunodefi-
ers of dead cells that are filled with keratin. Yet some vi- ciency virus (HIV).
ruses, for example, the papillomaviruses (Chapter 16),
can exploit cuts or abrasions in the skin to target the living
basal layer of the epidermis. Other viruses (e.g., yellow fe- THE INNATE IMMUNE SYSTEM
ver virus [Chapter 7]) take advantage of insects or other Should an invader succeed in breaching a physical barrier,
vectors, which penetrate the skin via a sting or bite. it then encounters the components of the innate immune
Since viruses cannot replicate in the outer dead layers system. This arm of immunity is so named because it is in
of the skin, most use the living epithelia of the respiratory, place to function nonspecifically against a broad range of
gastrointestinal, and urogenital tracts for their portals of invaders. Like the adaptive immune system, its compo-
entry into the host. In contrast to the multilayered skin, nents need to be activated by the invader. But since the
these epithelia are only one or a few cell layers thick. More- innate immune system does not need to tailor its response
over, they are actively transporting oxygen and nutrients to each particular invader, it can begin acting against the
to underlying tissue. In addition, the area covered by our invader days before the adaptive immune system is able to
skin is only about 2m2, in contrast to the 400m2 that are do so. Thus, the principal role of the innate immune re-
covered by the mucous membranes lining our respiratory, sponse is to provide an early, nonspecific, effective line of
gastrointestinal, and urogenital tracts. Thus, these epithe- defense against a broad range of potential invaders.
lia constitute a vulnerable portal to potential invaders. As noted above, the innate immune system recognizes
Yet because of their vulnerability, these epithelia have molecular patterns that are characteristic of broad classes
evolved formidable barriers of their own against infec- of microorganisms, doing so via receptors that are en-
tion. For example, in the case of the respiratory mucosa, coded in the germ line. In contrast, the adaptive immune
the mucous coating and the beating cilia combine to im- system recognizes antigenic determinants that are unique
pair virus binding to the cell surfaces and to sweep the to each particular invader, doing so via receptors that are
viruses up to the throat, where they may be coughed away not encoded in the germ line and that are generated by a
or swallowed. specific process, described below. When activated by a
As for the gastrointestinal tract, the low pH within the particular invader, the adaptive immune system tailors
gut provides a major block to infection by that route. In its response to that invader. However, the price for that
the stomach, the pH may be as low as 2.5, a level of acidity specificity is that activation can take 1 week or longer.
that is lethal to most viruses. Moreover, powerful prote- Since virus infectious cycles can be as short as 4 to 5h,
olytic enzymes such as pepsin provide yet another barrier the host does not have the luxury of doing nothing while
to infection via the gut. Should a virus somehow survive waiting for the adaptive immune system to respond ap-
passage through the stomach, it then encounters the di- propriately. Consequently, the innate immune system is
gestive enzymes of the small intestine, as well as the bile nearly entirely responsible for controlling virus infections
salts contributed by the liver. Together, the degradative ac- for the first 3days or so after a primary (first time) infec-
tivities of the enzymes and the detergent activity of the tion with a virus. (Recall that the adaptive immune system
bile salts make for a most formidable impediment against has a memory of previously encountered pathogens. This
any viral pathogen. memory enables adaptive immunity to already be in place
The epithelium of the vagina, like the skin, is several against previously encountered pathogens, even years af-
cell layers thick. Moreover, only the deepest, most inacces- ter the primary infection. Importantly, memory enables
sible basal cells of the vaginal epithelium are actively di- adaptive immunity to spring into full action much more
viding. Since viruses replicate best in cells that are actively quickly during reinfections than during primary infections.)
Host Defenses and Viral Countermeasures 83

Indeed, the hosts need to mount an immediate response innate immunity effectors than in the absence of adaptive
to a primary invasion almost certainly was a major selec- immunity effectors only. The effectiveness and crucial role
tive pressure behind the evolution of the innate immune of innate immunity are likely the reasons why vaccinia vi-
system. Yet the role of innate immunity does not end rus and the herpesviruses express a variety of proteins
there, since the innate immune response plays key roles in that are not essential for virus replication invitro but help
orchestrating a specific adaptive immune response when these viruses to evade innate immune responses in vivo.
the latter is needed. Such viral proteins impede interferon (IFN), complement,
The crucial role of the innate response as quarterback and inflammatory responses and oppose apoptosis (see be-
of the adaptive response is illustrated by the substantial low). The quarterback role of the innate response in activat-
impairment of immune protection in individuals lacking ing adaptive immunity is considered in more detail later.
innate immune function. Importantly, such individuals As noted above, the innate immune system recognizes
have vastly more severe infections than individuals who molecular patterns that are common to a variety of patho-
are deficient for adaptive immunity only (Fig. 4.1). Infec- gens but are not present on uninfected host tissues. The
tions go largely uncontrolled in individuals with inherited lipopolysaccharide components of the coats of gram-
deficiencies of innate immunity, because the adaptive im- negative bacteria provide a well-known example of a
mune response cannot be deployed in the absence of the common molecular pattern that is recognized by the in-
quarterback role of the innate response. In contrast, in in- nate immune system. In the case of viruses, the structures
dividuals who are deficient in adaptive immunity only, of both enveloped and nonenveloped viruses contain reg-
the innate response may be able to initially control the ularly repeating surface features that may be recognized
infection, although it likely will not be able to clear it. by the innate immune system. Moreover, as described be-
These findings may account for the rarity of individuals low, the innate immune system can also react to virus-
with inherited deficiencies in innate immune function. specific nucleic acids.
Experiments using knockout mice that were made de- When the innate system senses the presence of these
ficient for effectors of either innate immunity or adaptive foreign molecular patterns, it attempts to wall off, con-
immunity lead to the same conclusion. Mice infected with tain, and then eliminate the invading pathogen. To do so,
vaccinia virus (a poxvirus) or herpes simplex virus (HSV) the cells of the innate system respond by secreting a set of
undergo much more severe infections in the absence of chemical effectors called cytokines. (Cytokines are pro-
teins that are secreted by cells in order to mediate specific
effects on other cells that express specific cytokine recep-
Figure4.1 The crucial quarterback role of the innate response is tors.) These secreted cytokines trigger the inflammatory
illustrated by the relative effects of deficiencies in innate immunity response, which is exemplified by the swelling, redness,
versus deficiencies in adaptive immunity. Infections go largely uncon-
trolled in individuals with an inherited innate immunity dysfunction
and flu-like symptoms that are characteristic of many vi-
because an adaptive immune response cannot be effectively estab- rus infections. How the innate system knows what cyto
lished in the absence of the quarterback role of the innate immune kines to secrete and when to do so is as follows.
response. In contrast, in individuals who are deficient in adaptive Cells of the innate immune system, in particular mac-
immunity functions only, innate immunity may be able to initially
contain the infection, even if it cannot clear it. Adapted from P. Parham,
rophages, are present behind the physical barriers, where
The Immune System, 2nd ed. (Garland Science, New York, NY, 2005), they function as sentries, on the alert for invading patho-
with permission. gens. They express a family of receptors known as Toll-
like receptors (TLRs), each of which recognizes a particu-
Lacking innate
lar kind of molecular feature that is common to a broad
immunity only class of pathogens. For example, TLR4 at the cell surface
Number of microorganisms

binds to the lipopolysaccharides of gram-negative bacte-

ria. Together, the 10 known TLRs can recognize virtually
every pathogen that can cause infection. (Mammalian TLRs
Lacking adaptive
immunity only
were originally identified and studied in macrophages.
However, TLRs have also been found in nonmyeloid cells,
in which their role is not yet well clarified. Recently,
functional TLRs have been found in nonmyeloid human
cells invivo [e.g., gingival fibroblasts] and in a variety of
Healthy humans nonmyeloid human and mouse cell lines in culture.)
A number of viruses now are known to trigger innate
Duration of infection responses via TLRs. For example, respiratory syncytial

virus and mouse mammary tumor virus have been shown different players interact and synergize with each other to
to trigger an innate response through TLR4, whereas provide an impressive degree of effectiveness. Moreover,
measles virus and human cytomegalovirus (HCMV) ap- they likewise orchestrate and synergize with the mediators
pear to trigger an innate response via TLR2. Also, TLR3 of adaptive immunity (see below). We now consider the
and TLR8, which are expressed on intracellular membranes, individual players of innate immunity.
bind to double-stranded and single-stranded RNA mole-
cules, respectively, which are generated during a variety of Cytokines: the IFNs
virus infections (see below). Thus, at the very earliest stages In the late 1950s, A. Isaacs and J. Lindenmann observed
of the virus-cell interaction, TLRs seem to sense the virus that culture media from chicken cells exposed to inacti-
and activate an inflammatory response against it. vated influenza virus contained an activity that would
Once TLRs of a macrophage are engaged, they trans- render fresh uninfected cells more resistant to infection
mit a signal that induces it to release a particular set of with certain viruses. Moreover, this antiviral activity could
cytokines appropriate for dealing with the particular class be detected within hours of exposure to the virus. Subse-
of invader. For example, TLR3signals induce the expres- quent protein fractionation showed that the interfering
sion of IFNs, which are the major antiviral cytokines (see activity actually is comprised of a set of related proteins,
below). Cytokines also recruit additional innate immune now known as alpha-interferon (IFN-) and IFN- (also
effector cells to the infected site and induce inflammation referred to as type I IFNs). Thirteen related IFN- proteins
(Box4.1). are presently known.
The components of the innate immune response include Since type I IFNs could be induced by inactivated in-
the professional phagocytes (i.e., macrophages and neu- fluenza virus, virus replication per se is not required for
trophils), natural killer (NK) cells, and a variety of cyto their induction. What then induces these IFNs?
kines that these and other cells produce. The complement Double-stranded RNA produced inadvertently during
system of proteins too constitutes an effector mechanism virus replication is thought to trigger IFN induction, as
of innate immunity. Importantly, bear in mind that these implied by the experimental finding that type I IFNs can
be induced with synthetic double-stranded RNA (e.g.,
polyinosine/polycytosine; see below). Consequently, one
might expect that IFN is not specific for any particular vi-
Box 4.1 rus and that the IFN activity induced by one kind of virus
might have an antiviral effect against unrelated viruses.
The discovery of TLRs in the late 1990s helped to clarify Indeed, that is so. The antiviral state induced by type I IFN
the workings of the innate immune system. Moreover, is not virus specific. In this regard, recall that nonspeci-
their discovery resulted in a better appreciation of the ficity is a characteristic feature of innate immune re-
importance of innate immunity in host defenses. Al- sponses in general. However, IFN is host specific. For
though much was known earlier about the IFNs and NK example, human IFN induces an antiviral effect in hu-
cells (see the text), innate immunity still received rela- man cells but not in mouse cells.
tively little attention compared to the plentiful attention There is a second pathway of IFN- and IFN- induc-
lavished on the adaptive immune system, which seemed tion, which like the first is independent of virus replica-
intrinsically more glamorous with its specific antibody- tion but, in this instance, is even independent of virus
producing and cell-killing functions. gene expression. It is observed in so-called natural IFN-
producing cells (IPCs) of hematopoietic origin, the prin-
The TLRs derive their name from their resemblance
cipal one of which is a particular type of immature den-
to a fruit fly protein called Toll. Flies that lack Toll are
dritic cell (see below). Induction occurs when these IPCs
defective for their dorsal-ventral differentiation. The
are exposed to any of a variety of enveloped viruses, even
gene is called Toll, since that is the German word for
when those viruses have been physically or chemically in-
fantastic, although in this instance it may have the
connotation weird. Interestingly, Toll in flies also acts to
activated. Induction is thought to result from the direct
protect these insects from fungal infection. In addition, interaction of viral envelope glycoproteins with the sur-
plants too contain TLRs, one of which (called N protein) face of the IPC. (Proviso: Much of what is known about
protects tobacco against tobacco mosaic virus. Thus, Toll- the activities of the IFNs [and of other cytokines as well]
like proteins appear to have arisen early in evolution, and is known from studies of model systems in cell culture. Yet
they likely played a crucial role in making multicellular cell cultures cannot be expected to reflect the complexities
organisms possible. of the mixed cytokine responses that likely take place
invivo [see below]. Consequently, the actual workings and
Host Defenses and Viral Countermeasures 85

effects of these cytokines, as they occur invivo, are only For example, TLR3 of human corneal epithelial cells ex-
partly understood.) perimentally triggers production of IFN- in response to
Before continuing our discussion of the two type I the synthetic double-stranded RNA polyinosine/polycy-
IFNs, we should note that they, together with gamma IFN tosine. (Polyinosine/polycytosine also triggers the activa-
(IFN-) (a type II interferon), are among the most impor- tion of proinflammatory cytokines interleukin 6 [IL-6]
tant cytokines that are induced at early times in a variety and IL-8 by these cells.) Importantly, corneal epithelial
of viral infections. The type I IFNs are best known for cells are a first line of defense against ocular pathogens.
their antiviral activities, whereas the type II IFN, IFN-, is (As noted above, functional TLRs have been found in
best known as an important modulator of immune sys- mammalian cells other than macrophages and other cells
tem function. Yet the type I IFNs are also important mod- of the myeloid lineage. Human corneal epithelial cells are
ulators of immune system function, and IFN- also ex- an example of a cell type that is not of the myeloid lineage
presses an antiviral activity, but one that is different from but nevertheless expresses TLRs.)
that of the type I IFNs. Other TLRs can trigger expression of IFN-/. Fur-
For convenience, IFN- and IFN- are often referred thermore, recent experimental results show that the retin-
to collectively as IFN-/. Yet there are differences between oic acid-induced gene I product (RIG-I) also may be an
them. IFN- is produced in leukocytes, and IFN- is important intracellular sensor of double-stranded RNA
broadly produced in fibroblasts and epithelial cells. Almost in virus-infected cells. Indeed, RIG-I appears to be essen-
any cell will produce either IFN- or IFN- in response to tial for activating transcription of IFN-/ during infec-
being infected by a virus. In contrast, IFN- is produced by tion by several RNA viruses, including paramyxoviruses,
T cells of the adaptive immune system and by NK cells of flaviviruses, and influenza viruses. These experimental
the innate immune system, in each case doing so in re- findings were obtained using mouse knockout cells and
sponse to extracellular signals. The crucial importance of animals (see Box4.4).
the IFN responses is demonstrated by animals that cannot Next, we consider the means by which type I IFNs
express IFN defenses. These animals are more quickly bring about an antiviral state and the means by which that
killed by viruses, and at lower infectious doses than are re- antiviral state protects the host. We begin by noting that
quired to kill animals that express normal IFN responses. IFN-/ do not necessarily protect the virus-infected
Viruses are known to induce IFN-/ early after the cells in which they are first induced. Instead, the IFN mol-
onset of infection, as shown in chickens, mice, nonhuman ecules secreted by virus-infected cells have a more impor-
primates, and humans, and they presumably do so in tant effect on as yet uninfected neighboring cells, in
other avian and mammalian species as well. The mecha- which they actually induce an antiviral state. Upon
nisms for inducing IFN-/ are known only in part. The binding to specific receptors on these neighboring cells,
most potent known inducer of these IFNs is double- IFN-/ transmit a signal that results in the transcrip-
stranded RNA, which many different viruses generate tional up-regulation of about 300 genes (as demonstrated
during their replication. In the case of single-stranded by gene array technology). The expression of these genes
RNA viruses, IFN induction happens as follows. Single- establishes an antiviral state that prepares IFN-activated
stranded RNA viruses replicate by first generating single- cells to defend themselves against viral invaders. The up-
stranded RNA molecules that are complementary to the regulation of IFN-responsive genes occurs only when
genomic RNA. These complementary RNA molecules extracellular IFN binds to IFN-specific receptors on the
then serve as templates for the production of progeny vi- plasma membrane.
ral genomes. Usually, there is no double-stranded RNA IFN-/, as well as IFN- and other cytokines (e.g., in-
replicative intermediate as such (might this be an IFN eva- terleukins; see below), transmit signals through receptors
sion strategy?), but the complementary RNA molecules that activate the JAK/STAT signaling pathway (Fig. 4.2).
produced during replication may inadvertently anneal with The significance of JAK/STAT signaling is as follows. Recall
genomic RNA, forming double-stranded RNA molecules that mitogenic transmembrane signals, which are induced
that can induce IFN-/. In the case of DNA viruses, many by growth factors (e.g., the epidermal growth factor),
of these viruses transcribe both strands of certain regions generally are propagated via a series of protein kinases
of their genomes. Thus, these DNA viruses also produce that activate each other in a particular sequence, usually
RNA molecules with complementary sequences that beginning with a tyrosine kinase and followed by a suc-
inadvertently may anneal, thereby generating double- cession of serine/threonine kinases. In contrast, the JAK/
stranded RNA that can trigger IFN production. STAT signaling pathway is a short, direct pathway that en-
TLR3 recognizes and responds to double-stranded RNA, ables cells to respond quickly to cytokine stimulation. Its
as noted above, and it then triggers expression of IFN-/. key features are as follows.

IFN binding causes dimerization of the individual IFN
receptor chains, so that their associated JAKs can phos-
phorylate tyrosine residues in the cytoplasmic domains of
Plasma membrane the receptor polypeptides. Transcription proteins named
STATs (acronym for signal transduction and transcription
factors) bind to the phosphorylated receptors and in turn
are phosphorylated by the associated JAKs. The phospho-
rylated STATs dimerize and move into the nucleus, where
they bind to specific transcriptional control sequences
called IFN-stimulated response elements, and in so doing
P P they activate specific IFN-regulated target genes.
There are seven distinct STAT genes in mice. Inactivat-
P ing stat-1 causes the mouse to be unable to mount an in-
nate immune response against bacterial and viral infec-
tions. Inactivating stat-4 or stat-6 severely compromises
the ability of the mouse to mount an adaptive immune
P response. These experimental findings underscore the
STAT crucial roles played by the JAK/STAT signaling pathway in
both innate and adaptive immunity systems. As seen be-
low, a variety of cytokines signal via the JAK/STAT signal-
ing pathway.
Next, we see that IFNs induce a variety of innate anti-
viral processes.
Of the several hundred cellular genes that are upregu-
lated by IFN-/, three in particular are known to play
Nucleus important roles in the antiviral state. (It is a safe bet that
P more than a few others will also be found to play important
roles.) These three genes encode the RNA-activated pro-
P Transcription tein kinase PKR, RNA-activated RNase L, and the double-
stranded RNA-specific adenosine deaminase ADAR. We
consider each of these host gene products in turn.
PKR levels are upregulated 5- to 10-fold by IFN-/.
Yet, although PKR levels are up-regulated by IFN-/,
Figure4.2 Simplified depiction of the JAK/STAT signaling pathway, this serine/threonine protein kinase remains inactive in
by which IFN-/, as well as IFN- and other cytokines (e.g., interleu- uninfected cells. However, it is ready to be activated if
kins [see the text]), transmit signals. IFN binding causes dimerization
of the individual IFN receptor chains, enabling their associated JAKs
the cell should become infected. This occurs as follows.
to phosphorylate tyrosine residues in the cytoplasmic domains of the PKR has a double-stranded RNA-binding motif. When
receptor polypeptides. Transcription proteins named STATs bind to two molecules of PKR bind to the same molecule of
the phosphorylated receptors and in turn are phosphorylated by the double-stranded RNA, these two PKR molecules cross-
JAKs. The phosphorylated STATs dimerize and move into the nucleus,
where they bind to specific transcriptional control sequences called
phosphorylate each other, thereby activating each others
IFN-stimulated response elements, thereby activating specific target serine/threonine kinase activity. Thus, if a cell in the anti-
genes. Adapted from G. M. Cooper and R. E. Hausman, The Cell: a viral state is infected, the double-stranded RNA that is
Molecular Approach, 4th ed. (ASM Press, Washington, DC, 2007), with generated by the virus binds and activates PKR.
The activated PKR acts by phosphorylating numerous
target proteins, but the best-understood antivirus effect of
There is one cell surface receptor for IFN-/ and an- PKR involves eIF2a, an initiation factor for protein syn-
other for IFN-. The cytoplasmic domains of these recep- thesis. PKR phosphorylates eIF2a, thereby inactivating its
tors are associated with protein kinases of the Janus kinase ability to initiate translation, which brings about an abrupt
(JAK) family, named for Janus, a god of Roman mythol- shutdown of both viral and cellular protein synthesis in
ogy, who stood watch over gates and doorways. (Janus saw the infected cell. Since cellular protein synthesis is inhib-
in two directions at once and thus is depicted with two ited by the IFN-activated PKR, the cell may be sacrificed
faces.) in the process. However, this is usually preferable to
Host Defenses and Viral Countermeasures 87

having it live to support the replication and dissemination Yet another role for IFN-/ is to have an effect on NK
of the virus. cells. As described below, NK cells are ready to kill virus-
RNase L is a nuclease that degrades most viral and infected cells without having to first be activated. However,
cellular RNAs. Like PKR, it is activated by a multistep NK cells become more efficient killers of virus-infected
process. First, IFN-/ induces a 10- to 1,000-fold up- cells when they are activated by signals alerting them that
regulation of RNase L production, as well as that of an- an infection is under way. IFN-/ are among the most po-
other cellular enzyme, 2-5-oligo(A) synthetase. But as in tent of these NK cell activators, enhancing the base level of
the case of PKR, these enzymes remain inactive until the NK cell cytotoxicity 20- to 100-fold. This example is one of
cell might be infected. Should the cell be infected, the several that illustrate how two or more players of the innate
double-stranded RNA that is generated by the virus acti- immune response synergize to strengthen innate defenses.
vates the 2-5-oligo(A) synthetase, which then produces Next, we consider the induction and activities of IFN-.
2,5 oligomers of adenylic acid. This unusual polynucle- Recall that IFN-/ may be induced in almost any cell
otide then binds to, and activates, RNase L, which goes on that is infected by a virus. In contrast, IFN- is produced
to degrade most viral and cellular RNAs, thereby shutting and secreted nearly exclusively by T cells and NK cells.
down viral replication. Since cellular RNAs too are de- T cells are key effectors of adaptive immunity. They gener-
graded by activated RNase L, this enzyme, like PKR, may ate IFN- only after their specific immune receptors recog-
kill the cell while blocking viral replication. nize their cognate antigenic determinants (i.e., the specific
ADAR, a double-stranded RNA-specific adenosine antigenic determinants that a particular T cells or B cells
deaminase, catalyzes the deamination of adenosine to receptor recognizes and binds to), which are presented to
inosine on viral and cellular RNAs. This site-specific RNA them via a special process, described later. In contrast, NK
editing results in the translation of defective proteins. cells, which are effectors of innate immunity, are induced
Still another IFN-induced antiviral effect, in this in- to produce IFN- by the IFN-/ given off by virus-infected
stance resulting from long-term exposure to IFN-/, in- cells, as well as by the cytokines TNF- and IL-12. The last
volves the induction of a number of cellular genes that two cytokines are secreted by macrophages in a positive-
lead to programmed cell death (apoptosis). Whereas this feedback loop with NK cells, as described in the next para-
leads to the early suicidal death of infected cells, thereby graph. Like IFN-/, IFN- transmits signals to target cells
blocking virus production within them, it also leads to the by binding to a heterodimeric receptor, which activates the
suicide of nearby uninfected cells. This seemingly extreme JAK/STAT signaling pathway (Fig. 4.2). IFN- receptors are
response benefits the host, since (as argued above) block- present on the plasma membranes of most cell types.
ing replication of a viral pathogen is well worth the loss of IFN- mediates a variety of antiviral and immune reg-
a relatively few cells. Moreover, these cells might have been ulatory effects, involving both innate and adaptive im-
killed anyway by the virus, but only after producing many mune functions. The following example involves NK cells
progeny viruses. (Note that apoptosis is a normal host of the innate immune system. IFN- secreted by NK cells
process, which serves several important functions not re- can activate macrophages. These activated macrophages
lated to immune defenses. For example, apoptosis elimi- then secrete TNF- and IL-12. (The designation IL is for
nates cells in which the DNA might be damaged, in that interleukin. Interleukins are cytokines that mediate com-
way eliminating cells that might go on to become cancer- munication between leukocytes.) TNF- and IL-12, which
ous. It also plays an important role in development. A are thus produced by the activated macrophages, in turn
well-known example is in the elimination of tissue to gen- stimulate the NK cells to produce more IFN-, in that way
erate the spaces between the fingers and toes.) establishing a positive-feedback loop between the NK cells
IFN-/ also may cause an inflammatory response. and macrophages (Fig. 4.3). The net result is that these
Inflammation is discussed in more detail below in the crucial immune effector cells are primed to efficiently
context of tumor necrosis factor alpha (TNF-). For carry out their immunity functions.
now, note that this effect of IFN-/ can lead to the ap- One of the best-characterized effects of IFN- on
pearance of common flu-like symptoms, such as fever, macrophages is the induction of the enzyme nitric oxide
chills, muscle aches, nausea, and malaise. Indeed, these synthetase 2 (NOS2). NOS2 generates nitric oxide (NO),
symptoms, which are typically experienced during an which in turn reacts with oxygen to yield other reactive
influenza virus infection, result from the inflammatory molecules. Together, these molecules can inactivate es-
response rather than from the direct action of the virus sential virus proteins. A variety of important human
per se. This explains why a variety of viruses, despite be- pathogenic viruses (e.g., HSV, Epstein-Barr virus [EBV],
ing unrelated to influenza virus, cause a similar set of hepatitis B virus [HBV], and vaccinia virus) were shown
flu-like symptoms. to be sensitive to the action of NOS2. Indeed, experiments

Virus by affecting the local blood capillaries, thereby enabling

Virus- fluid, plasma proteins, and cells to pass from the circula-
tion to the infected tissue. Inflammation is a critically
important antiviral response. It is essential to initiate anti-
IFN-/ viral defenses, while also contributing significantly to the
overt symptoms that result from infection.
NK IL-2 TNF- is produced initially by macrophages at the in-
fected site. These cells, which are described in more detail
Make Proliferate below, are stationed under the physical barriers and in the
Make more IFN- IL-12
tissues as well. They might be thought of as both immune
IFN- IL-12
sentries and first responders. Tissue macrophages may be
TNF- able to deal with the few viruses that penetrate the physi-
cal barriers. However, if the tissue macrophages cannot
Macrophage handle the situation, they call for reinforcements. They do
so by secreting TNF-. Importantly, macrophages pro-
duce inflammatory levels of TNF- only when there is a
Figure4.3 IFN-/ given off by virus-infected cells induces NK cells sustained engagement of a number of macrophages, an
to produce IFN-. The IFN- secreted by the NK cells in turn activates event which occurs only when the infection is outpacing
macrophages. The activated macrophages then secrete TNF- and the macrophages at the infected site. Thus, inflammation
IL-12. TNF- and IL-12 thus produced by the macrophages in turn
occurs only when local innate defenses are seriously en-
stimulate NK cells to produce more IFN-, in that way establishing a
positive-feedback loop between the NK cells and the macrophages. gaged and in danger of being overrun.
The net result is that these crucial immune effector cells are each Neutrophils are the major leukocytes of the inflamma-
primed to efficiently carry out their functions. tory response. They leave the blood and hone in on the
infected site in response to TNF- (and also cytokines
IL-1 and IL-8), which is released by the embattled mac-
with mice showed that impairing NO production results rophages. Neutrophils and macrophages are described in
in uncontrolled infection by a variety of viruses. more detail below. However, it is important to note here
IFN- also influences B cells and T cells of the adaptive the following distinction between these cell types. Al-
immune system. How it does so is explained in The though both macrophages and neutrophils are profes-
Adaptive Immune System, below. Thus, do not worry if sional phagocytes, macrophages are stationed in the tissues,
the remainder of this paragraph is not yet comprehensi- whereas neutrophils remain in the blood circulation, un-
ble. The material discussed therein will come up again less recruited to an infected site. The reason neutrophils
later. IFN- helps induce B cells to undergo antibody class are normally retained in the circulation is that they are
switching to produce immunoglobulin G (IgG) antibod- professional killers. When they move from the blood to
ies, which are good at opsonizing virus and binding com- the surrounding tissues, they are immediately ready to kill
plement. IFN- also affects T cells, doing so indirectly by target cells. Consequently, they have the potential to do
inducing macrophages to secrete IL-12. This cytokine in much unintended damage to uninfected host tissue, and
turn induces T cells to secrete a cytokine profile that is it is therefore important that they do not leave the circula-
especially effective against bacterial and viral infections, as tion, unless they are needed to contain an infection. Neu-
opposed to defending against a parasite. trophils make up about 70% of the white blood cells in
the circulation.
Cytokines: TNF-a, Some Other Cytokines, For the reason just given, the exit of neutrophils from
and Inflammation the blood circulation is carefully regulated. Indeed, exit of
TNF-a regulates immune defenses against viruses as well neutrophils from the circulation requires multiple signals
as against other pathogens. Although it affects immune from the infected site. Moreover, these signals induce local
defenses in a variety of different ways, the major role of changes both in circulating neutrophils and in the local
TNF- is to heighten inflammation. Inflammation is a capillary endothelial cells. Each of these changes is nec-
rather general term, referring to the local accumulation of essary to trigger extravasation, the movement of cells
fluid, plasma proteins, and leukocytes (mainly neutrophils from the blood circulation to the surrounding tissue. We
and NK cells) at the infected site. This accumulation consider the steps of neutrophil extravasation in the next
causes the familiar swelling, reddening, heat, and pain as- paragraph. Note the multiple roles that TNF- plays in
sociated with inflammation. TNF- induces these changes the process.
Host Defenses and Viral Countermeasures 89

Both TNF- and IL-1 are secreted by activated mac- adhesion molecules on leukocytes and capillary en-
rophages at the infected site. Together, they induce the ex- dothelial cells. Similarly, IFN- and IL-12 cooperate with
pression of adhesion molecules on the luminal plasma TNF- to affect the activities of leukocytes at the infected
membranes of local capillary endothelial cells. As neutro- sites.
phils (and other leukocytes as well) circulate through
those local capillaries, TNF- and IL-1 also induce the ex- Macrophages, Neutrophils, NK Cells, and
pression of adhesion molecules on these neutrophils. The Antibody-Dependent Cellular Cytotoxicity
upregulated adhesion molecules on the neutrophils and Macrophages are large professional phagocytes that are
on the luminal surfaces of the capillary endothelial cells stationed in the tissues beneath the mucosa, positioned so
act together to cause the neutrophils to slow down near as to engulf viruses or bacteria that may have succeeded in
the infected site and roll along the luminal surface of the penetrating the mucosa. Note that other macrophages are
capillaries. IL-8, which also is secreted by activated local stationed in the lymph nodes, where they are positioned
macrophages, induces conformational changes in these to trap and degrade pathogens that arrive in the lymph
upregulated adhesion molecules, which has the effect of from the sites of infection. In the lymph nodes, these mac-
strengthening the adhesions between the neutrophils and rophages may play an important role in the activation of
capillary endothelial cells. The net result is that the neu- adaptive immunity. That is, they present antigens de-
trophils stop rolling, thereby enabling them to respond to rived from the trapped pathogens to T cells, as explained
yet other alarm signals coming from the infected site. These in The Adaptive Immune System, below.
additional signals prompt the neutrophils to cross the Above, we discussed the positive-feedback loop be-
capillary endothelia to the infected region, by a process tween macrophages and NK cells, in which TNF- and
known as diapedesis. TNF-, produced at the infected IL-12 secreted by macrophages activate NK cells, while
site, also affects the activities of neutrophils and other IFN- from NK cells activates macrophages (Fig. 4.3). In-
leukocytes appearing there. Moreover, TNF- induces deed, macrophage activation results primarily from the
vascular endothelial cells to make platelet-activating fac- influence of IFN-. IFN- also plays a role in the activation
tor, which causes blood clotting and blockage of local of the helper T lymphocytes and cytotoxic T lymphocytes
blood vessels. This prevents pathogens at the infected site (CTLs) of the adaptive immune response (an example of
from entering into the circulation and using it as a con- the interaction between innate and adaptive immunity;
duit to disseminate the infection. see below). Nevertheless, macrophages are the main tar-
In addition to the local effects caused by cytokines at gets of IFN- activity. (Lymphocytes are collectively a
an infected site, cytokines also induce distant effects that class of white blood cells that includes the B lymphocytes
also heighten host defenses. For example, TNF-, IL-1, [B cells] and T cells of the adaptive immune system and
and IL-6 elevate body temperature, giving rise to fever, by the NK cells of the innate immune system. Importantly,
acting on temperature control sites in the hypothalamus. B cells and T cells have a large number of variable receptors,
They also act on muscle and fat cells, altering their me- each of which is specific for a particular antigen, while NK
tabolism, thereby causing them to generate heat. These cells have a small fixed number of receptors against general
cytokines are referred to as endogenous pyrogens, to dis- molecular features present on a variety of pathogens.)
tinguish them from the exogenous pyrogens produced by Many studies have shown that IFN- enhances the abil-
certain bacterial pathogens, which also cause fever. ity of macrophages to destroy ingested bacterial and pro-
Although we can attest to the fact that fever can cause tozoan pathogens. Yet the importance of IFN- in the in-
considerable discomfort, its purpose is to enhance the nate immune response against viruses appears to depend
ability of the immune system to fight the infection. Fever on the particular virus. For example, ifn-g-deficient mice
enhances host defense because most microbial pathogens show enhanced susceptibility to vaccinia virus and lym-
(including viruses) replicate less efficiently at elevated phocytic choriomeningitis virus (LCMV), while their sus-
body temperatures. Moreover, adaptive immune responses ceptibility to vesicular stomatitis virus and Semliki Forest
are improved at elevated temperatures. virus is unimpaired. Note that one can compare ifn-g
We noted above that IFN-/ may trigger apoptosis. knockout mice with ifn-a/b knockouts, as well as with
TNF- likewise may trigger apoptosis when it binds to its double knockouts, with respect to their susceptibility to
receptors on virus-infected cells. vaccinia virus and LCMV. The results show that mice in
Once again, note that these cytokines generally work in which the two IFN types are knocked out are more suscep-
various combinations to generate their effects. For exam- tible to these viruses than are single knockouts, showing
ple, activated macrophages produce IL-1 as well as TNF-. that the two IFN types are nonredundant and comple-
These cytokines act together to induce expression of mentary in their effects.

Note that activated T cells (see below) as well as NK antibody binds to its cognate antigen via the Fab region, the
cells produce IFN-, and either of these cell types can ac- Fc region remains free. Neutrophils, macrophages, and NK
tivate macrophages, but only NK cells can do so at early cells all have receptors on their surfaces that can bind to the
times. Indeed, the early production of IFN- by NK cells free Fc regions of antigen-bound antibody molecules. Thus,
makes these cells a central player in the regulation of the antibodies against viral glycoproteins can act as a bridge,
early stages of the immune response, especially regarding linking the above immune effector cells to virus-infected
intracellular pathogens. target cells that display viral glycoproteins at their sur-
Usually, local tissue macrophages can cope with the few faces. Neutrophils contain granules (actually modified
viruses that might penetrate the mucosa. In those instances lysosomes), which sequester hydrolytic degradative en-
where the macrophages might be overwhelmed by the in- zymes, nitric oxide, and toxic oxygen radicals. The con-
vader, help is provided by another professional phagocyte, tents of these granules are released by contact of the neu-
the neutrophil. As noted above, neutrophils comprise trophil with the target cell, leading to the target cells
about 70% of all circulating white blood cells. They leave demise. The purpose of this lethal assault on the infected
the blood and hone in on the infected site in response to the cell is to terminate virus replication within it.
cytokines IL-1 and TNF-, which are released by the em- Bear in mind that in ADCC, killing is carried out by
battled macrophages. In point of fact, neutrophils are the nonspecific neutrophils of the innate immune system.
major leukocyte in the inflammatory response (see above). However, antibodies that are produced by the adaptive
They too release cytokines that can influence subsequent immune response make ADCC specific against virus-
events, in particular, Mip-1, which is a chemoattractant infected cells. Thus, ADCC offers an example of the in-
for T cells and B cells of the adaptive immune response. nate and adaptive immune systems working in concert to
Neutrophils have been well studied as key components of contain an infection.
the innate immune response against bacteria and fungi. Al- Since ADCC requires the participation of specific anti-
though they are less studied in the context of virus infections, bodies, it cannot occur early during a primary infection by
there is evidence that neutrophils do help to control infec- the invader (i.e., the adaptive immune system needs to be
tions by some important viral pathogens, including HSV and activated to produce antibodies, a process that takes sev-
influenza viruses. Still, while neutrophils are known to be a eral days or more). Yet ADCC might play a role in control-
major component of inflammatory infiltrates and to secrete ling human infections involving HIV and herpesviruses.
cytokines that promote inflammation (and vascular damage (Why might ADCC be effective in these instances? Have a
as well) and activate adaptive immunity, there is relatively look back at Chapter 3.) Also, we might imagine that
little information available concerning how they might di- ADCC can provide an early defense during a second en-
rectly defend against viruses. Neutrophils are known to counter with a virus, in which case preformed antibodies
phagocytose bacteria, which they then degrade via an assort- already are present (see below).
ment of degradative enzymes. However, during virus infec- Another means by which neutrophils, and macrophages
tions, neutrophils generally do not persist at viral lesions too, are thought to contribute to antiviral defenses is by
but are replaced there by macrophages and lymphocytes. phagocytosing extracellular viruses. This occurs as follows.
One conceivable mechanism by which neutrophils might First, antibodies also bind directly to free virus particles,
help contain viral infections is referred to as antibody- as well as to viral glycoproteins in the plasma membranes
dependent cellular cytotoxicity (ADCC), a mechanism of infected cells. When an extracellular virus is decorated
that might be particularly effective against enveloped with bound antibodies, it is readily recognized by the Fc
viruses. ADCC works as follows. Recall that the envelope receptors on professional phagocytes, thereby enhancing
glycoproteins of many enveloped viruses are exposed at phagocytosis of the invader. A pathogen so decorated by
the host cell plasma membrane prior to virus maturation antibodies is said to be opsonized. When the pathogen has
and budding (Chapter 2). Consequently, these membrane- been phagocytosed, the endocytic vesicle that contains the
associated viral glycoproteins can be recognized by virus- pathogen fuses with a granule, and the invader is de-
specific antibody molecules, which are generated by the stroyed. Fc receptors on macrophages enable these cells to
adaptive immune response. Now, note the following fea- phagocytose and digest whole virus-infected cells that dis-
ture of antibody molecules (discussed in greater detail play viral antigens on their surfaces (Box4.2).
later in the chapter). The two antigen-binding domains of The granules of neutrophils also contain antimicrobial
an antibody molecule, which constitute the so-called peptides called defensins, which are 29 to 34 amino acids
Fab region, are at the two amino termini of the molecule, long. Defensins are active invitro against bacteria, fungi,
while the single invariant portion, referred to as the Fc and various viruses, and they presumably act in vivo as
region, is at its single carboxy terminus (Fig. 4.4). When an well. Viruses that are sensitive to defensins invitro include
Host Defenses and Viral Countermeasures 91

N termini Heavy N termini Antigen-

chain binding

chain Disulfide bonds
Constant Hinge region
Carbohydrate region

C termini
Figure4.4 IgG antibodies, which are the most common antibody class in the blood, are often
used to illustrate common structural features of antibodies. (A) Each IgG molecule is comprised
of two identical heavy chains (green) and two identical light chains (yellow). The two heavy
chains are linked by disulfide bonds, and each contains a carbohydrate moiety (blue). (B) The
antigen-binding sites are located at the N termini of the variable (V) region (red), which is
comprised of sequences on both the heavy and light chains. These sequences vary from one
IgG molecule to another and confer the specificity for a particular antigen. The constant region
(blue), which is the same for all antibodies in the class, is likewise comprised of sequences on
the heavy and light chains. IgG antibodies contain a flexible hinge region at about the middle of
the two heavy chains. The protease papain experimentally cleaves IgG molecules at the hinge re-
gion to generate three fragments: two Fab fragments and one Fc fragment. The Fab fragments
contain the antigen-binding sites on the antibody, whereas the Fc fragment provides the link to
other immune effectors. For simplicity, the carbohydrate is not shown here. Adapted from
P. Parham, The Immune System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.

HSV types I and II, CMV, type A influenza viruses, and macrophages have TLRs that recognize a variety of mo-
HIV. In contrast, echovirus and rhinovirus are defensin lecular features common to broad classes of pathogens,
resistant. Each of the defensin-sensitive viruses noted above and these receptors transmit signals that induce mac-
is enveloped, whereas the defensin-resistant examples are rophages to secrete cytokines. IL-12, which is one of the
nonenveloped. This raises the possibility that defensins at- important proinflammatory cytokines produced by TLR-
tack viral lipid envelopes. Much remains to be learned stimulated macrophages, activates NK cells to secrete IFN-.
about the defensins, as these versatile peptides appear to The IFN- produced by the NK cells in turn induces the
have other important immune activities, including serving macrophages to secrete TNF-. Macrophages have recep-
as chemoattractants for phagocytes and T lymphocytes. tors for TNF- as well as IFN-, and the combination of
The primary functions of NK cells are (i) to kill virus- these two cytokines hyperactivates the macrophages to se-
infected cells and (ii) to produce cytokines. In point of crete even more IL-12. NK cells have specific receptors for
fact, NK cells are the major producers of IFN- early in an both IL-12 and TNF-, and the combination of these cytok-
infection. Like B cells and T cells of the adaptive immune ines produced by the hyperactivated macrophages causes
system, NK cells are white blood cells or leukocytes. How- the NK cells in turn to produce even more IFN-. Thus, the
ever, NK cells are key players in the innate immune system positive-feedback loop noted above is established between
response to virus infections, whereas T cells and B cells are macrophages and NK cells at the infected site (Fig. 4.3).
the principal cells of adaptive immunity. To fulfill their func- Importantly, synthesis of IFN- early during infection de-
tions in adaptive immunity, T cells and B cells rearrange and pends upon the production of IL-12 by macrophages. Now,
express particular gene clusters to generate antigen-specific also recall that the ability of NK cells to kill virus-infected
cell surface receptors and soluble antibody molecules. These cells is stimulated 20- to 100-fold by IFN-/ produced by
developments are explained in detail later. In contrast, as virus-infected cells. Thus, stimulation of NK cells with IFN-
effectors of innate immunity, NK cells do not rearrange or / favors induction of NK cytotoxicity, whereas stimulation
express genes that might enable antigen-specific responses. of NK cells with IL-12 favors production of IFN-.
Recalling that a primary function of NK cells is to se- NK cells first were recognized for their ability to kill
crete cytokines, in particular IFN-, NK cells are induced virus-infected target cells, with the important qualification
to do so by macrophages as follows. As discussed above, that they did not need to be activated to do so. This may

Box 4.2 The rationale for NK cells preferentially killing cells

that display low levels of MHC class I molecules is that
many viruses attempt to evade CTL-mediated attack by
As you are likely aware, antibodies also may neutralize
down-regulating host cell MHC class I expression. NK cells
viruses. Neutralizing antibodies are discussed in The
Adaptive Immune System, below. Yet it is worth noting
counter this viral immune evasion strategy by attacking
here that antibodies alone, in actuality, do not kill any- and destroying cells that do not display sufficient levels of
thing. Instead, the role of antibodies is to bind to a spe- MHC class I molecules to deliver the inhibitory signal to
cific target. The role played by antibodies in ADCC and NK cells (Fig. 4.5). Many tumor cells likewise bear low lev-
in opsonizing viruses underscores the fact that antibody els of class I MHC molecules and are destroyed by NK cells.
binding marks a target for destruction by other immune (Proviso: The concept that NK cells kill virus-infected
effectors. In this regard, note that antibodies also tag cells because of MHC down-regulation is satisfying and
viruses for destruction by the classical pathway indeed is a widely accepted dogma. Nevertheless, there is
of complement activation, as described in the text. not a lot of experimental evidence to substantiate it.)
Once NK cells have identified a virus-infected target,
What then of neutralizing antibodies? First, only some how then do they administer the lethal blow? The answer
of the antibodies that bind to a virus actually neutralize is that NK cells induce infected target cells to commit sui-
it, that is, actually prevent it from initiating an infec- cide; that is, NK cells induce target cells to undergo apop-
tion. Second, the mechanisms by which neutralizing tosis. Their modes of inducing apoptosis are similar to
antibodies block infection are not well understood inall those used by CTLs (see below). In one such mode, the
instances. When some neutralizing antibodies bind to a adhering killer cell releases a mixture of perforin (a mem-
virus, they impair the ability of the virus to attach to the brane active protein similar to the complement protein C9
surface of a cell. However, other neutralizing antibodies [see below]) and the enzyme granzyme B onto the surface
do not prevent entry of the virus into the cell. of the target cell. The target cell then takes up these ligands
by endocytosis. Within the target cell, the perforin mole-
cules then create pores in the membranes of the endocytic
vesicles. This allows the granzyme B to enter the cytoplasm,
seem to contradict the preceding paragraph. But note that where it induces a chemical chain reaction leading to apop-
NK cells in fact exist in several stages of readiness to kill. tosis. (Note that other descriptions of this process have the
Resting NK cells are competent to kill. However, they released perforin acting at the plasma membrane of the tar-
become much more efficient killers when activated by get cell, thereby creating pores that enable the granzyme to
IFN-/. At any rate, the ability of resting NK cells to enter the target cell, where it induces apoptosis.) In another
kill virus-infected cells is in marked contrast to CTLs of pathway, NK cells use the Fas ligand (FasL) on their sur-
the adaptive immune system, which need to be activated faces to transmit an apoptotic signal into the target cell via
by a rather complex sequence of events (see below) before the Fas protein on the target cell surface. These apoptotic
they can kill virus-infected cells. pathways are described in greater detail later.
Since NK cells, as components of innate immunity, NK cells, like neutrophils, have Fc receptors. Thus, NK
do not have receptors for specific viruses, how do they cells likewise may also kill virus-infected cells by ADCC.
recognize virus-infected cells? Current understanding is As in the case of neutrophil-mediated ADCC, antiviral
that NK cells use two sets of receptors for this purpose. One antibodies bound to the infected cell trigger the NK cell
set comprises the activation receptors, about which little is attack. That is, antibodies against the viral glycoproteins in
known. The other set, the inhibitory receptors, recognize the host cell plasma membrane can act as a bridge between
major histocompatibility complex (MHC) class I molecules. the NK cell and the target cell. Moreover, when the NK cells
(The role of MHC class I molecules is described in detail be- Fc receptors bind to the antibody molecules on the target
low. Briefly, these proteins are present on the surfaces of vir- cell, the NK cell actually becomes a more effective killer.
tually all host cells. They are used by infected cells to display Also like neutrophils, NK cells may have the potential
antigenic peptides that are produced during infection by to do much damage in the tissues. Thus, again like neutro-
intracellular pathogens. CTLs have antigen-specific recep- phils, NK cells migrate from the circulation to infected
tors that recognize the combination of antigenic peptide sites only in response to multiple signals mediated by in-
and MHC molecule at the surface of the infected cell.) NK flammatory cytokines released from the infected site.
cells can attack and destroy only those cells that do not ac- And they exit from the blood circulation by the cytokine-
tivate the inhibitory receptors. Thus, NK cells preferentially triggered roll, stop process, by which neutrophils like-
kill cells displaying low levels of MHC class I molecules. wise leave the blood circulation (Box4.3).
Host Defenses and Viral Countermeasures 93

Interaction of NK cell with Interaction of NK cell with target cell

uninfected healthy cell in which MHC class I expression is lost

Lytic granules

NK cell NK cell

receptor Killing
MHC Ligand
class I

Healthy cell Virus-infected cell

(no MHC class I)

No killing of healthy cell Killing of virus-infected cell in which

MHC class I expression is inhibited
Figure4.5 A possible mechanism by which NK cells recognize virus-infected cells. NK cells use two sets of receptors
to determine whether to attack a target cell. One set comprises the activation receptors, about which little is known.
The other set, the inhibitory receptors, recognizes MHC class I molecules. NK cells attack and destroy only those cells
that do not activate the inhibitory receptors. Thus, NK cells preferentially kill cells displaying low levels of MHC class I
molecules. The rationale for this is that many viruses attempt to evade CTL-mediated attack by down-regulating host
cell MHC class I expression. NK cells provide a host countermeasure to this viral evasion strategy by destroying cells
infected by these viruses, since these cells do not deliver the inhibitory signal to NK cells. Adapted from P. Parham, The
Immune System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.

Since NK cells and neutrophils receive activating sig-

nals when their Fc receptors bind antibodies, and since
Box 4.3
activated NK cells and neutrophils can cause serious dam-
At several points in this chapter, we encounter the prin-
age to tissues, it is important that Fc receptors on these
ciple that the launching of certain potent weapons of the
cells be able to distinguish between antibody molecules that immune system requires multiple signals. This is remi-
are bound to a target cell and the majority of free antibody niscent of dual-key activation, in which two individuals
molecules. How might this be accomplished? A probable must act in concert to launch a ballistic missile. Such
answer is that activation of the killing response of these ef- a fail-safe system is installed to safeguard against the
fector cells requires a clustering of their Fc receptors, which inappropriate launching of a highly destructive weapon.
occurs only when they bind to an antibody-decorated Leave it to Mother Nature to have invented fail-safe
target cell. Alternatively, or in addition, the binding of systems well before Dr. Strangelove.
Fc receptors to antibodies is a low-affinity interaction.
However, the effector cell can bind to the antibody-coated
infected cell with high affinity. That is so because many vi-
ral envelope proteins accumulate in the target cell plasma
membrane, to which the Fab arms of many antiviral anti- While there are numerous studies demonstrating the
bodies might bind. And many Fc receptors are present on efficacy of ADCC in cell culture, there is not a lot of ex-
the NK cells and neutrophils, which can bind to the Fc re- perimental evidence that might support a role for this
gions of those target-cell-bound antiviral antibodies and, process invivo. However, the following example, involv-
in that way, be cross-linked. ing HSV, provides a tantalizing suggestion that ADCC

indeed is relevant in vivo, while also showing why it is nervous tissue in vivo. Moreover, bipolar bridging cannot
frustratingly difficult to prove it is so. be evaluated in the mouse model, since gE-gI does not
HSV and other herpesviruses produce envelope glyco- bind well to the Fc region of mouse antibodies. This is an
proteins that can bind the Fc portion of human antibody important obstacle, since the mouse model is particularly
molecules. This envelope glycoprotein in the case of HSV valuable because it is an invivo system in which it is pos-
is called gE-gI. It protects HSV-infected cells from ADCC sible to generate knockouts (Box 4.4). (If gE-gI did react
in vitro, presumably as follows. The Fab arms of anti- with the Fc regions of mouse antibodies, then what sort of
HSV-specific antibodies bind to the other HSV envelope experiments might you carry out in the mouse model to
glycoproteins (i.e., gC and gD) in the host cell plasma demonstrate the invivo efficacy of bipolar bridging on the
membrane. But these antibodies simultaneously bind to part of the virus and of ADCC on the part of the host?)
gE-gI in the plasma membrane via their Fc region. This Despite uncertainty concerning a role for ADCC in
phenomenon, known as bipolar bridging, prevents the Fc controlling virus infections in vivo, there is ample evi-
regions of the cell-bound anti-HSV antibodies from en- dence for the protective role of NK cells, irrespective of
gaging the Fc receptors of neutrophils and NK cells. The their role in ADCC. For example, patients who lack NK
example continues as follows. cells suffer from severe persistent virus infections, in par-
Since the herpesvirus gE-gI protein is present in the ticular, infections involving HSV and CMV (also a herpes-
viral envelope, this protein also protects the virus from virus). Yet those rare NK cell-deficient individuals indeed
destruction via the classical complement pathway (see remain able to mount normal adaptive immune re-
below). Thus, bipolar bridging is likely a general strategy sponses, demonstrating the importance of NK cells in the
that herpesviruses evolved to evade host antibody- immune response to at least some virus infections. In ad-
mediated defenses. Yet it is difficult to study the signifi- dition, NK cell depletion and reconstitution experiments
cance of bipolar bridging invivo. For example, gE-gI also with mice also demonstrate the importance of NK cells in
mediates the spread of the virus in both epithelia and controlling at least some virus infections invivo (Box4.4).

Box 4.4
Much of our thinking about the human immune response Also useful are mice with severe combined immunodefi-
to microbial pathogens invivo is based on experiments ciency disease (SCID). SCID mice are homozygous for a
using animal models. These model systems indeed have rare recessive allele that prevents them from making an en-
provided important insights into the multifaceted mam- zyme necessary for assembling the DNA sequences encoding
malian immune response to virus infections. Nevertheless, the variable regions of antibody molecules and T-cell recep-
bear in mind that although a human virus might infect an tors [i.e., they are defective for V(D)J recombination; see the
animal model, the features of the infection in the animal text]. Consequently, these mice fail to develop both humoral
(e.g., portal of entry, dissemination, sites of replication, and and cellular adaptive immune responses, and thus, they are
pathogenesis) might not entirely reflect the pattern of events largely helpless against viruses. Since B cells and T cells fail
in the human patient. Furthermore, there are differences be- to develop in these mice, they accept cell and tissue grafts
tween the immune systems of different mammalian species, from other strains of mice and even from humans, mak-
and different host species may use different components of ing them a useful model for modeling the human immune
the immune system to combat the same virus. For these rea- system invivo. SCID mice are used experimentally to assess
sons, a particular animal model is not likely to reflect all the the importance invivo of particular effectors of immunity
important features of the corresponding infection in humans. (which may be added by passive transfer) against particular
Nevertheless, despite these stipulations, animal models have pathogens.
provided singularly important insights into the workings of
mammalian immunity and of viral countermeasures. Another mouse model involves thymectomized and congen-
itally athymic or nude mice, in which T-cell development
The most useful experimental animal models have involved and adaptive immunity can be restored with thymic tissue
mice. Among these are models involving highly inbred grafts. The ability to generate transgenic mice and knock-
strains, in which all mice are of the same MHC haplotype, out mice also provides important opportunities to evaluate
thereby allowing experiments involving the adaptive trans- specific immune effectors. See also Box4.6.
fer of T cells, B cells, and antibodies.
Host Defenses and Viral Countermeasures 95

Moreover, increased NK cell-mediated cytotoxicity al- Despite the above, the actual significance of the com-
ways follows induction of IFN-/ invivo, implying that plement system in the human response to virus infections
the response plays a physiologically important role. Fur- is not known with certainty. This is so in part because
thermore, NK cells also produce an important group of complement deficiency conditions in humans have thus
cytokines, including IFN- and TNF-, that restrict virus far been associated only with recurrent or chronic bacte-
infections in multiple ways. Finally, the relevance of NK rial infections. However, mice depleted of complement
cells in the control of at least some virus infections is un- components were shown to have more severe infections
derscored by the various strategies that viruses have with Sindbis and influenza viruses. Moreover, early in
evolved to evade NK cell-mediated defenses (see below). infection, before the appearance of virus-specific anti-
body, complement in some way controls the viremia in
The Complement System complement-expressing animals. Another good reason
The complement system is a key component of innate im- for believing that complement plays a significant role in
munity. It is comprised of about 30 different proteins that the control of virus infections is that pathogenic viruses
are synthesized in the liver and circulate in the plasma. have evolved a variety of strategies to evade complement
Several of these proteins have protease activity, but they activation, as described later in the chapter.
are released in an inactive form, known as zymogens. Activation of the complement system can occur via
Thus, like other immune system players, the complement several pathways. The first of these is the classical path-
system needs to be activated. As described below, activa- way, so named because it was the first complement activa-
tion involves a cascade of proteases, in which each enzyme tion pathway to be discovered. Activation via the classical
cleaves and activates the next enzyme in the pathway. complement pathway is dependent on antibodies. This is
When activated, the complement components work in contrast to the alternative complement pathway, which
together to carry out several distinct functions. First, these is antibody independent.
proteins cooperate to destroy some viruses, as well as vi- Activation of the classical complement pathway occurs
rus-infected cells. They do this by forming a membrane as follows. Complement protein C1 binds to the Fc region
attack complex (MAC) that can destroy viral envelopes of either IgG or IgM antibodies that are bound to the target
and punch holes in the membranes of infected cells. Sec- (Fig. 4.6A). (There are four main classes of antibodies: IgM,
ond, the deposition of complement on viruses may neu- IgG, IgA, and IgE. They differ in their Fc regions, and those
tralize them, consistent with the concept that neutralization differences make each class particularly well suited to carry
of viruses can be explained by the progressive accumula- out its particular tasks. IgM and IgG antibodies are superb
tion of exogenous protein on the virus surface. (As de- and good, respectively, at fixing complement. The roles of
scribed below, viruses can bind complement via both an the different antibody classes are discussed below, in The
antibody-dependent and an antibody-independent path- Adaptive Immunity System.) C1molecules contain a ser-
way, and either of these pathways can result in virus neu- ine protease activity that is activated when C1 binds to the
tralization. Neutralization via complement alone may be antibody. The activated C1 then sequentially cleaves com-
important in the early phase of a primary infection, be- plement proteins C4 and C2, to generate C4b and C2a,
fore antibody is present.) Third, phagocytes have recep- respectively (Fig. 4.6B). Some of the C4b molecules that
tors for complement. Consequently, complement can op- are generated bind covalently to the target membrane,
sonize pathogens, thereby promoting their destruction by where they interact with soluble C2b to form membrane-
phagocytosis. Finally, fragments of the complement pro- bound C4b2a. Importantly, C4b2a serves as a C3 con-
teins that are generated during its other functions also vertase, cleaving C3 to generate C3a and C3b (Fig. 4.6C).
serve as chemoattractants, which recruit other immune The cleavage of C3into C3a and C3b is the key event
system players to the infected site. in the activation of the complement system by both the
The crucial importance of the complement system is classical and alternative pathways. The reasons are as fol-
demonstrated by the fact that the rare humans who are lows. C3b binds covalently to proteins and carbohydrate
born with a defect in one of the complement proteins groups on the pathogen surface, a step referred to as
usually do not survive infancy. Yet despite the rarity of hu- complement fixation. C3b is then able to bring together
mans with complement deficiencies, defects in a variety of the proteins that constitute the MAC. (C3b is also the
different complement components have been observed. complement product that opsonizes the pathogen for de-
Among these are defects in the C3 complement compo- struction by phagocytosis.) First, C3b forms a complex
nent, which occupies an important place in both the with C4b2a, designated C4b2a3b (Fig. 4.6C), which cleaves
classical (antibody-dependent) and alternative (antibody- C5 to generate membrane-associated C5b and the free
independent) complement pathways (see below). fragment C5a (Fig. 4.6D). The last four components of

C5a C8 C9
C1 C3a C6 C7
C4 C2 C3 C5
a b a b
Antibody C3b

Antigen C4b 2a C4b 2a 3b C5b C5b67

Cell surface

Membrane C5b6789
attack complex

Figure4.6 The classical pathway of complement activation. (A) Complement protein C1 binds to the Fc region of
either IgG or IgM antibodies that are bound to the target. (B) C1molecules contain a serine protease activity that is
activated when C1 binds to the antibody. The activated C1 then sequentially cleaves complement proteins C4 and C2,
to generate a complex designated C4b2a, which binds to the cell surface. (C) C4b2a then activates C3. That is, C4b2a
serves as a C3 convertase, cleaving C3 to generate C3a and C3b, the latter of which binds to proteins and carbohy-
drate groups on the cell surface, forming a complex with C4b2a, designated C4b2a3b. (D) C4b2a3b then cleaves C5 to
generate membrane-associated C5b and the free fragment C5a. (E) The last four components of the complement
cascade, C6, C7, C8, and C9, then bind to the cell surface to form the MAC that lyses the pathogen.

the complement cascade, C6, C7, C8, and C9, then bind to antibody for its activation. Instead, it depends on the slow
the target surface to form the MAC (Fig. 4.6E), which ly- spontaneous cleavage of C3, the most abundant comple-
ses the pathogen by forming pores in its membrane. Thus, ment protein. This spontaneous cleavage of C3 generates
the lesions generated by these MACs have the same effects a C3b-like molecule that binds to membranes in a fashion
as those generated by perforin released by CTLs and NK similar to the C3b molecules generated by the classical
cells. Actually, the C9 complement protein is homologous pathway (Fig. 4.6). As in the classical pathway, the sponta-
to perforin. neously generated C3b binds B, which is next activated by
Several other points about the classical pathway should factor D to form Bb, which is then bound in the C3bBb
be noted. First, the C4b2a complex is active only briefly. complex. Importantly, C3bBb is the C3 convertase of the
But in that time it generates 1,000 or so C3b molecules, alternative complement pathway. A positive-feedback loop
many of which are membrane bound. These membrane- is thereby established, generating more C3b and C3bBb,
bound C3b molecules can bind complement factor B, which greatly accelerates the generation of MACs (Box4.5).
which is then cleaved by complement factor D to yield the If the C3b-like protein of the alternative complement
Bb fragment, which forms a complex with C3b known as pathway does not react with a target membrane in about
C3bBb. Importantly, C3bBb, like C4b2a, has C3 convertase 60 s (microseconds), it is inactivated by two plasma pro-
activity. Thus, the number of C3b molecules deposited at teins: complement factor I (CFI) and complement factor H
the site of complement fixation, and hence the number (CFH). This inactivation of C3b is important to maintain
of MACs that assemble there, is much greater than the functionality of the complement system. For example, in
number of C4b2a complexes. patients with factor I deficiency, C3b accumulates un-
Second, since the classical pathway for complement ac- checked, leading to the accumulation of the C3 convertase,
tivation is dependent on antibodies, it provides an excel- C3bBb. The end result is that the reservoir of C3 in the
lent example of the cooperation between effectors of the blood is depleted, and factor-I-deficient individuals are
innate and adaptive immune systems. Moreover, it under- thus more susceptible to a variety of infections.
scores the fact that antibodies by themselves do not kill Has the following occurred to you? Since the alterna-
pathogens. Rather, they mark their targets for destruction tive pathway for complement activation occurs spontane-
by other immune effectors. In fact, the complement pro- ously, and since C3b binds nonspecifically to proteins and
teins are so named because they complement the antigen- carbohydrate groups on membranes, what is there to pre-
binding function of antibodies. vent that nasty molecule from initiating attack on our own
The alternative pathway for complement activation dif- healthy cells? In point of fact, since these complement
fers from the classical pathway in that it does not require proteins can be devastating if inappropriately activated on
Host Defenses and Viral Countermeasures 97

Box 4.5
Here are a couple of unexpected developments. First, apropos virus, and mumps virus. The explanation for these seemingly
the classical complement pathway, several viruses, includ- paradoxical observations is not clear, despite the fact that
ing retroviruses, Sindbis virus, and Newcastle disease virus, this phenomenon was first observed more than 25years ago.
activate the classical complement pathway in the complete Moreover, it is not even clear how virus-infected cells are able
absence of antibody. Second, apropos the alternative comple- to activate the alternative complement (i.e., the antibody-
ment pathway, although that pathway is activated in the independent) pathway, although it presumably involves some
absence of antibody, in some instances the biological activity effect of the virus on the plasma membrane. As you may have
of the alternative pathway is augmented by specific antibody. supposed, there also are reports of virus-infected cells being
For example, in the case of measles virus-infected HeLa cells, lysed by the classical pathway.
although the alternative complement pathway is activated in
the absence of antibody, infected cells are not lysed unless IgG While noting the gaps in our understanding of comple-
antibodies also are present. In an important control experi- ment-mediated lysis of virus-infected cells via the alter-
ment, lysis of these measles virus-infected cells was not di- native pathway, we should add that its significance as an
minished by the absence of complement components specific antiviral measure invivo also is not known with certainty.
to the classical (i.e., the antibody-dependent) complement However, it well may be an important antiviral defense
pathway. Thus, the effect of antibody in this example does not invivo, since it occurs early to an infected cell, before the
involve the classical pathway, and the role of antibody in this release of progeny virus. Moreover, it might occur early in
lytic process is unknown. Similar experimental results were the course of infection, before the adaptive immune system
obtained in studies of cells infected by HIV, HSV, influenza can be activated to clear the infection.

healthy cells, there are about as many proteins that regu- Since MBL binds to carbohydrate moieties found on
late complement activation as there are proteins facilitat- common pathogens, the lectin activation pathway is more
ing it. One such regulatory protein is the decay-accelerating specific in its targeting of an invader than the alternative
factor (DAF; also called CD55). DAF is present on the sur- pathway, which is indiscriminate. Yet the lectin activation
faces of human cells, where it accelerates the breakdown pathway is less specific than the classical pathway, in which
of C3b, as well as C4b of the classical pathway. This action antibodies direct complement activation against specific
by DAF prevents formation of both the alternative and pathogens. In any event, in the lectin activation pathway,
classical C3 convertases. The host membrane proteins MBL has a function similar to, but less specific than, that
protectin (CD59) and membrane cofactor protein (CD46) of antibodies in the classical pathway. Also, note that MBL
have similar activities. is one of a family of proteins known as collectins, all of
Before moving on, we should note that there is yet a which are soluble pattern recognition receptors. All inter-
third pathway for complement activation: the lectin ac- act with glycoproteins and carbohydrates on microbial in-
tivation pathway, which is known to be effective in the vaders. Some cause microorganisms to aggregate, or they
defense against bacterial and yeast pathogens that have may opsonize microorganisms to increase their phagocy-
mannose-containing glycoproteins exposed on their sur- tosis. They are considered a first line of defense against
faces. Importantly, there are no mannose-containing car- pathogens that display appropriate residues. MBL is pres-
bohydrates on human cells. The key player in this pathway ent only at low levels in the absence of infection. Together
is the protein mannose-binding lectin (MBL; a lectin is a with the complement proteins, it is one of the so-called
protein that can bind to a carbohydrate). MBL circulates in acute-phase proteins that are produced by the liver after
a complex with an inactive enzyme called MBL-associated an inflammatory infection. Their generation is stimulated
serine protease, which is activated when MBL binds to a by the cytokines produced at the site of infection.
target. Activated MBL-associated serine protease cleaves As we have noted, neutrophils and macrophages bear
C4 and C2 (much as C1 cleaves C4 and C2in the classical surface opsonic receptors for complement. Consequently,
pathway) to initiate complement activation. Note that complement, like antibodies, may facilitate the destruc-
MBL also was shown to bind to surface carbohydrates of tion of pathogens by phagocytosis. However, while op-
viruses such as HIV, HBV, and influenza virus, implying sonization by antibodies must await the induction of
that complement activation via the lectin pathway may adaptive immunity, complement deposition via the alter-
help to regulate infections by these viruses. native pathway occurs spontaneously, and opsonization

by complement can thus occur early during a primary in- protein prevents PKR activation by competing for bind-
fection. Opsonization by antibodies can yet play an early ing to double-stranded RNA. (The reovirus replication
role during a second encounter with a pathogen. cycle is covered in detail in Chapter 5. For now, try to con-
Have you been wondering what role, if any, might be ceive of stratagems by which reoviruses might conceal the
played by the free C3a, C4a, and C5a fragments that are replication of their double-stranded RNA genomes from
generated during activation of the complement pathways? the host cell. Later, you can ask whether reoviruses actu-
These fragments do not bind to the invader. Instead, each ally use one of your stratagems.)
functions as a chemoattractant that recruits inflamma- Influenza viruses have single-stranded RNA genomes.
tory cells to infected sites. For example, C5a is a chemoat- Yet, like other viruses that have single-stranded RNA
tractant that targets the transport of neutrophils from the genomes, influenza virus may inadvertently form
blood circulation to an infected site, as discussed above. double-stranded RNA during infection. (As noted above,
Moreover, at the infected site these complement protein single-stranded RNA viruses generally do not replicate via
fragments activate macrophages and neutrophils to be double-stranded RNA replicative intermediates. Still,
more effective killers. Thus, the complement system makes double-stranded RNA may inadvertently form if a viral
efficient use of its various products. RNA were to hybridize with its RNA template postsynthe-
So, we have the complement proteins being able to de- sis.) Yet influenza virus is somewhat effective at counter-
stroy viruses by (i) building MACs, (ii) facilitating the ing the action of IFN-/ by producing a protein, NS1,
phagocytic destruction of the pathogen, and (iii) enhanc- which impairs several pathways of interferon induction
ing the inflammatory response to control the infection. and blocks several IFN effectors as well (Chapter 12). For
Moreover, the complement proteins may also (iv) neutral- example, NS1 binds to double-stranded RNA to prevent
ize some viruses and (v) destroy virus-infected cells. Impor- it from activating PKR (Table4.1).
tantly, the complement system can do these things very fast. Here is an immune evasion strategy that I find particu-
There is yet one other function of the versatile comple- larly intriguing. Some viruses generate soluble homologs
ment system that merits noting. Antibodies generated by of IFN receptors, which are secreted from infected cells, so
adaptive immunity interact with the invader to produce a that they might inhibit extracellular IFN from signaling
potentially pathologic accumulation of antigen/antibody neighboring cells. That is, these virus-encoded IFN recep-
complexes in lymphoid organs and kidneys. The comple- tor homologs act as decoy receptors that intercept extra-
ment system acts on these immune complexes in several cellular IFN molecules, not unlike sending out decoys as a
ways. First, it opsonizes the larger complexes, thereby fa- countermeasure to an antiballistic missile system. Exam-
cilitating their elimination by phagocytosis. Second, the ples of this stratagem are provided by vaccinia virus and
fixing of complement also leads to the breakdown of the myxoma virus, a poxvirus of rabbits (Table4.1).
immune complexes by a mechanism referred to as the Another viral anti-IFN strategy is to block the JAK/
detergent-like effect. STAT signaling pathway, which mediates signaling via the
IFN receptor. JAK/STAT signaling can be impaired by in-
creasing the degradation rate of JAK/STAT or by blocking
Viral Evasion of Innate Immunity STAT phosphorylation. Some viruses that are known to
Evasion of IFNs employ these stratagems are listed in Table4.1.
Since IFN-/ and IFN- mediate a variety of crucial early Another strategy is to block the activities of IFN-in-
host antiviral defenses, we might expect that viruses have duced effectors, such as RNase L and PKR. For example,
evolved a variety of countermeasures to elude these IFN- hepatitis C virus (a flavivirus) blocks the activity of PKR
based defenses. Indeed, viral countermeasures are known by producing two proteins, E2 and NS5A. E2 acts as a de-
that act against every step in the sequence of IFN-related coy substrate for PKR, whereas NS5A forms heterodimers
events, from IFN induction to activation and expression with PKR to inhibit its activity.
of effector function (Table4.1). The poxviruses, which are the largest of the animal vi-
The reoviruses have evolved an especially remarkable ruses, display the most extensive array of countermeasures
stratagem to avoid inducing IFN-/. Despite the facts against IFN. For example, vaccinia virus secretes homologs
that reoviruses have double-stranded RNA genomes and of IFN receptors, as noted above. Remarkably, poxvirus-
that double-stranded RNA is the major inducer of IFNs, encoded IFN receptor homologs include proteins that
reoviruses nevertheless manage to go through their entire bind to each of the three IFNs (i.e., IFN-, -, and -), ef-
replicative cycle without ever generating nonencapsulated fectively neutralizing their activity at the infected site.
free double-stranded RNA that might otherwise induce Moreover, vaccinia virus produces a protein, A18R, which
IFN expression. And, for good measure, the reovirus 3 acts to prevent aberrant transcription of opposite DNA
Host Defenses and Viral Countermeasures 99

Table 4.1 Viral anti-IFN strategies

Strategy Virus examplesa Mechanism
Secreted homologs of vIFN-a/b receptors B18R (VV) Inhibits extracellular IFN-a/b
Secreted homologs of vIFN-g receptors B8R (VV) Inhibits extracellular IFN-g
M-T7 (MYX)
Intracellular homologs of vIRF vIRF-1/K9 (HHV-8) Represses IFN-inducible genes
vIRF-2 (HHV-8)
vIRF-3/K10.5 (HHV-8)
Inhibition of PKR dsRNA binding E3L (VV) Competes with PKR for binding to
s3 (Reo) dsRNA
NSP3 (Rota)
dsRNA-like analogs VA1-RNA (Ad) Binds and inactivates PKR
Substrate modifications or competition K3L (VV) Homolog of eIF2a
ICP 34.5 (HSV) Dephosphorylation of eIF2a
Enzyme inhibition US11 (HSV) Blocks PKR activity
PK2 (Bac)
Tat (HIV)
E2 (HCV)
Inhibition of RNase L ? (HSV) Synthesis of 2-5 oligoadenylate analogs
? (HIV, EMCV) Induction of RNAse L inhibitors
Translation protection g34.5 (HSV-l) Interferes with translational shutoff
Inhibition of IFN-induced gene expression ElA (Ad) Represses IFN-specific transcriptional
TP (HBV) responses
NS1 (IV)
Interference with JAK/STAT pathway ElA (Ad) Inhibits signaling downstream of IFN
T-Ag (Py) receptors (e.g., JAK/STAT)
V (SV5)
Tax (HTLV)
gp55 (SFFV)
E7 (HPV-16)
Reo, reovirus; Rota, rotavirus; IV, influenza virus; Bac, baculovirus; EMCV, encephalomyocarditis virus; Py, polyomavirus; SV5, simian virus 5; HBV, hepatitis B virus;
HTLV, human T-cell leukemia virus; SFFV, spleen focus-forming virus; VV, vaccinia virus; myx, myxoma virus; HHV-8, human herpesvirus 8; Ad, adenovirus; EBV,
Epstein-Barr virus; HIV, human immunodeficiency virus; HSV, herpes simplex virus; HCV, hepatitis C virus; HCMV, human cytomegalovirus. From D. C. Johnson and
G. McFadden, Viral immune evasion. In S. H. E. Kaufmann, A. Sher, and R. Ahmed (ed.), Immunology of Infectious Diseases (ASM Press, Washington, DC, 2002), with

strands that might then give rise to double-stranded RNA. successful viruses invivo. The explanation for this seem-
For good measure, this virus also expresses proteins that ing paradox is not yet known.
directly block activation of PKR and of RNase L (Box4.6). Recalling that the IFNs and NK cells can activate cel-
Together, the multifaceted ways in which IFNs can lular apoptotic pathways, we might expect that viruses
impair virus replication and the variety of anti-IFN coun- have evolved countermeasures to deal with the premature
termeasures that viruses have evolved underscore the cell death induced by those effectors of innate immunity.
importance of IFNs in host defense. Yet there are viruses That indeed is the case. But since the destruction of in-
that do not display any known anti-IFN countermeasures. fected cells by CTLs, which are key effectors of adaptive
Moreover, replication of these viruses (e.g., vesicular immunity, is of paramount importance for the clearance
stomatitis virus) in animals may provoke IFN responses, of a virus infection, and since the induction of apoptosis is
and these viruses indeed may be severely sensitive to the sole means by which CTLs destroy virus-infected cells,
IFN treatment in cell culture. Nevertheless, they remain we will take up viral countermeasures against apoptosis
100 CHAP TER 4

Box 4.6
The variety of countermeasures that viruses have evolved with the mutant virus, infection was found to be attenuated
to counter IFNs, and indeed to counter other host defenses relative to infection with wild-type virus. Since attenuation
as well, is rather remarkable. However, these viral coun- might very well have resulted from some reason unrelated
termeasures are known for the most part from studies in to the mutants inability to countermand PKR, such as im-
cell culture. Thus, it is important to consider what sort of paired replication of the mutant virus, the experiment was
experimental approaches might verify that a particular pu- repeated in mice in which the PKR gene was knocked out.
tative viral countermeasure indeed plays a role invivo. In these PKR knockout mice, the mutant virus displayed
wild-type levels of virulence (Leib etal., 2000). The com-
The following example, involving the HSV ICP34.5 gene, bined use of virus genetic manipulation and knockout mice,
which blocks PKR (Table4.1), demonstrates that this activ- as illustrated in this example, is a powerful general approach
ity of ICP34.5indeed plays a role during infection invivo. to evaluating possible virus virulence factors.
Moreover, it provides an example of a valuable experimental
approach for evaluating the importance of putative virus Reference
virulence factors. The first experimental step was to delete Leib, D. A., M. A. Machalek, B. R. G. Williams, R. H. Silverman, and
H. W. Virgin. 2000. Specific phenotypic restoration of an attenuated
the ICP34.5 gene from the viral genome. Second, normal virus by knockout of a host resistance gene. Proc. Natl. Acad. Sci.
mice were infected with either the engineered mutant virus USA 97:60976101.
or the wild-type virus. When normal mice were infected

after we have considered the effector mechanisms of CTLs it is believed by some investigators that they might pro-
(see below). mote the development of malignancies in the HCMV and
HHV-8 examples.)
Evasion of Cytokines In contrast to the viroreceptors that are expressed at
Cytokines other than IFNs, such as the chemokines, which the cell surface, the secreted cytokine-binding proteins
guide lymphocytes to sites of infection, and the interleu- constitute a second major class of cytokine modulators.
kins, which are produced by leukocytes, also play key roles The myxoma virus (a poxvirus) T2 protein, which mimics
in activating and regulating antiviral immune defenses. the binding domain of the TNF receptor, was the first of
Thus, it is not surprising that viruses have evolved multi- these soluble decoys to be discovered. In this instance,
ple strategies to impair and subvert cytokines in general. myxoma virus apparently appropriated the segment of
Moreover, since the IFNs are a subset of the cytokines, the TNF receptor gene that encodes that proteins ligand-
some of the strategies used by viruses against IFNs are binding domain, while leaving behind the gene segment
employed against other cytokines as well. For instance, that encodes the transmembrane anchoring sequence.
herpesviruses and poxviruses generate homologs of cy- Thus, myxoma virus is able to produce a soluble protein
tokines and cytokine receptors as a cytokine evasion that impedes TNF from interacting with the infected cell,
strategy. A sampling of some of the proteins encoded by in that way protecting the cell from TNF-induced lysis.
poxviruses to evade host immunity, including molecular The herpesvirus saimiri ECRF3 protein, which binds IL-8,
homologs of cytokines and cytokine receptors, are depicted provides an example of a soluble cytokine-binding pro-
in Fig. 4.7. Some specific examples of viral anticytokine tein encoded by a herpesvirus.
factors and stratagems are as follows. Another viral anticytokine strategy that employs mo-
Members of the herpesvirus and poxvirus families en- lecular mimicry is for viruses to generate cytokine ho-
code cytokine receptor homologs (viroreceptors) that are mologs, which act as antagonists (i.e., nonsignaling com-
expressed at the cell surface. For example, the K2R protein petitors) of the cellular ligand. The MC148R protein of
of swinepox virus is a decoy receptor for IL-8, thereby im- molluscum contagiosum virus (MCV) (a poxvirus) is one
peding binding of the cytokine to its cellular receptors. of the best-studied examples of a viral cytokine antago-
(Interestingly, the US28 protein of HCMV [a herpesvirus] nist. This chemokine homolog inhibits the binding of both
and the ORF74 protein of human herpesvirus 8 [HHV-8], CXC and CC chemokines to their respective receptors (see
also known as the Kaposis sarcoma herpesvirus, are vi- Chapter 21 regarding CXC and CC chemokine receptors
roreceptors that actually transmit signals in response to and the roles they play during HIV infection). MC148R is
host chemokines. The functional significance of these able to act as an antagonist, rather than as an agonist, be-
viroreceptor-mediated signals remains controversial, but cause it lacks the N-terminal region of chemokines, which
Host Defenses and Viral Countermeasures 101

Block proinflammatory
and immunoregulatory
activity of host IFN-

TNF receptor
IL-1 homolog IL-18 binding
MFG proteins
IFN- receptor
VGF (C11R) homolog VCP
Chemokine Complement
homolog binding
IFN- receptor Poxvirus DNA
homolog Cytoplasm CrmA
Viral gene expression
Nucleus Serpins

IL-8 receptor homolog

Binds IL-2 GIF Chemokine binding protein
IL-1 receptor homolog SPV-K2R
vCKBP Serpin 1

Binds to CC B28R
chemokine Inhibits plasmin, urokinase-type
plasminogen activator, tissue-type
plasminogen activator, thrombin,
and factor Xa
Figure4.7 Some of the proteins encoded by poxviruses to evade or subvert the host immune system. Molecular
mimicry of cytokines and cytokine receptors is a common immune evasion strategy adopted by members of the herpes
virus and poxvirus families of large DNA viruses. Note the soluble proteins that are secreted from infected cells, which
function as decoy receptors for IFN-, TNF, and various cytokines (including chemokines and interleukins). Poxvi-
ruses also secrete homologs of humoral immune regulators (virokines), such as the viral IL-10, vascular endothelial
growth factor (VGF), and myxoma viral growth factor (MGF), which is a homolog of epidermal growth factor (EGF).
The last two factors may stimulate neoplastic cell proliferation. Pathogenic poxviruses also encode a complement con-
trol protein. In fact, the vaccinia virus complement control protein (VCP) was one of the first soluble microbial pro-
teins thought to have a role in immune evasion. The figure does not depict the activities of a particular poxvirus but
instead is a composite of activities detected during infections of several different poxviruses, including vaccinia virus,
cowpox virus, molluscum contagiosum virus, and myxoma viruses. Adapted from P. Jha and G. J. Kotwal, J. Biosci.
28:265271, 2003, with permission.

is involved in receptor activation. The observation that in- A fourth anticytokine strategy used by a variety of vi-
flammatory cells are generally absent from MCV lesions ruses is to impair cytokine-mediated signal transduction
invivo, despite extensive virus replication at infected sites, downstream of the cytokine receptors by interfering with
supports the premise that the antagonistic effect of the JAK/STAT pathway (Table4.1). Finally, some viruses
MC148R provides an effective MCV chemokine evasion impair the secretion of proinflammatory cytokines by
stratagem. blocking the activating enzyme caspase-1. Note that
On the other hand, some secreted viral cytokine ho- caspase-1 also plays a key role in activating apoptosis, and
mologs, such as the UL146 protein of HCMV, are agonists the viral anti-caspase-1 activity may work to block apop-
(i.e., proteins or peptides that activate a receptor), rather tosis as well (see below).
than antagonists, of chemokine receptors. The result in
the HCMV example is that more leukocytes are recruited Evasion of NK Cells and ADCC
to the infected site, where they may facilitate dissemina- The human and murine cytomegaloviruses are known to
tion of the infection. The HIV Tat protein likewise acts as employ an ingeniously diabolical strategy to defeat NK cell-
a chemoattractant, at least invitro. mediated defenses. These members of the herpesvirus
102 CHAP TER 4

family actually produce an MHC class I molecule look- these effects may help HIV-infected cells to evade both
alike, which engages the NK cell inhibitory receptor to NK cells and CTLs and to establish chronic infection.
generate a dont kill signal.
What evidence might there be that the abilities of these Evasion of Complement
herpesviruses to produce an MHC class Ihomolog, and Viruses have evolved a variety of strategies to evade
thereby evade NK cells, might be functionally relevant complement-mediated defenses. For example, both HIV
invivo? Consider the following. All of the known herpes- (Box4.7) and vaccinia virus incorporate host cell CD46,
viruses have the ability to give rise to lifelong persistent CD55, and CD59into their envelopes when they bud from
infections. Indeed, this is a hallmark of the family. In con- the host cell plasma membrane. As noted above, these
trast, among the poxviruses, MCV is thus far unique for host proteins are present in cellular plasma membranes to
its ability to establish persistent infections. Moreover, prevent the complement system from initiating attack on
MCV also is unique among poxviruses in its ability to ex- healthy host cells.
press an MHC class I homolog. Thus, the persistent virus HCMV not only incorporates CD46 and CD55 into
lifestyle in some cases may depend on the ability of the vi- progeny virus particles but also up-regulates their expres-
rus to evade NK-cell-mediated lysis. (Also, recall that HSV sion at the plasma membrane. Thus, free virus is protected
and other herpesviruses produce membrane glycoproteins against complement and, moreover, virus-infected cells
that bind antibody Fc regions, thereby thwarting ADCC.) are protected against lysis by antibody plus complement.
HIV is a champion among viruses at evading host im- Vaccinia virus and the related variola and cowpox viruses
munity. As such, HIV has several mechanisms for down- display yet another countermeasure against complement.
regulating expression of MHC class I molecules on in-
fected cells, in that way enabling the virus to evade attack
by CTLs of the adaptive immune system. That fact and the
ability of HIV to persist throughout the 10years or more
Box 4.7
of an HIV infection (Chapter 21) might well lead you to HIV may very well be the most diabolical of all viruses,
expect that HIV also is able to impede NK-cell-mediated as it is able to resist virtually all of the immune defenses
cytolysis of HIV-infected cells. Indeed it is, and its overall that humans are able to mount against it. Regarding HIV
strategy in that regard is rather remarkable. The HIV Nef and complement, activation of the complement system
protein selectively down-regulates the human MHC class I indeed leads to the deposition of complement on free
genes HLA-A and HLA-B, while not affecting HLA-C. (The HIV particles and on HIV-infected cells as well. Yet HIV
group of genes encoding the human MHC proteins is re- infection is nearly entirely unaffected by complement,
ferred to as the human leukocyte antigen [HLA] complex even when anti-HIV antibodies are present. As noted
because MHC proteins are found on leukocytes [white in the main text, HIV particles incorporate CD55, the
blood cells]). Although HLA-C proteins can present viral complement decay-accelerating factor, at budding. Yet
peptides to CTLs, they are the predominant molecules CD55 is only partly responsible for the resistance of HIV
that transmit inhibitory signals to NK cells, thereby pre- to complement. Another HIV countermeasure against
venting killing of a target cell by an NK cell. The net result complement results from the ability of the HIV envelope
of the HIV strategy is that the MHC class I peptide-pre- glycoproteins gp120 and gp41 to interact with comple-
sentation pathway is sufficiently compromised in HIV- ment factor H (CFH), a humoral negative regulator of
infected cells to impair attack by CTLs, while the unim- complement activation (see the text). This interaction is
very likely relevant invivo, as shown by the following. Al-
paired expression of HLA-C enables HIV-infected cells to
though HIV-infected cells are ordinarily resistant to com-
also resist attack by NK cells.
plement, when infected cells are cultured in medium that
There also is a report that HIV up-regulates expression
is depleted of CFH, the addition of anti-HIV antibodies
of the nonclassical MHC class I molecule HLA-E on the
plus complement results in cell destruction. In contrast,
surface of lymphocytes. HLA-E molecules have limited uninfected control cells are not affected by the addition
sequence variability and do not play a role in antigen pre- of anti-HIV antibodies plus complement. These experi-
sentation, as do HLA-A, HLA-B, and HLA-C molecules. mental findings demonstrate that the ability of HIV-
Instead, HLA-E is believed to negatively regulate NK cell infected cells to resist the effects of complement depends
activity by delivering an inhibitory signal to NK cells. primarily on the ability of the HIV envelope glycoproteins
Thus, the HIV-induced up-regulation of HLA-E expres- to associate with CFH. Interestingly, the streptococci have
sion may further enable HIV-infected cells to evade attack evolved a similar strategy for resisting complement; the
by NK cells, despite the fact that the virus down-regulates streptococcal M proteinalso binds CFH.
expression of HLA-A and HLA-B molecules. Together,
Host Defenses and Viral Countermeasures 103

These viruses encode a protein that is similar to the hu- studies of how Vif promotes HIV infection of T cells re-
man serum complement control protein C4, which ac- vealed a new intracellular innate defense mechanism, as
celerates the breakdown of C3b and C4b and conse- well as a new viral evasion strategy, as manifested by Vif.
quently the decay of both the alternative and classical C3 T cells express a protein known as APOBEC3G, which is a
convertases. cytidine deaminase that converts cytidines to uridines.
HSV likewise encodes a protein, gC, which accelerates The unabated activity of APOBEC3G during HIV infec-
decay of the alternate C3 convertase. In addition, the HSV tion would hypermutate the HIV genome, thereby ren-
gC protein also blocks the interaction of C5 with C3 dering it noninfectious. Vif binding to APOBEC3G targets
(Fig. 4.6). Together, these effects of gC protect both free that host protein for degradation by sending it to the pro-
virus and HSV-infected cells from complement. (In mice, teasome. (The actual state of affairs may be more com
HSV gC mutants produce less severe skin lesions than plicated, since APOBEC3G can restrict HIV, even when
does wild-type virus, consistent with the premise that missing its cytidine deaminase activity, and Vif-induced
complement would act to contain HSV infection invivo, degradation of APOBEC3G may not be the only means by
were it not for the action of gC. However, since the HSV which Vif counteracts that protein [Chapter 21].)
gC proteinalso functions as a virus attachment protein, It is not yet known whether APOBEC3G has any other
how might one ascertain that the less severe pathology of role in the cell. Also, it is not known how many other vi-
the gC mutants results from an inability to impair com- ruses APOBEC3G might act against. However, APOBEC3G
plement action rather than from less efficient virus at- appears to be a major factor that impedes HBV from es-
tachment? Answer: C3 knockout mice were used. In these tablishing a chronic infection. In this regard, HBV, like
knockout mice, HSV gC mutant virus and wild-type virus HIV, makes use of reverse transcription during its replica-
were equally pathogenic. Therefore, gC can mediate resis- tion. (Interestingly, an enzymatic activity similar to that
tance to complement invivo [Box4.6].) of APOBEC3G is responsible for initiating the process of
Antibodies are properly thought of as crucial effectors somatic hypermutation, which introduces mutations into
of adaptive immunity. Yet antibodies also interact with specific regions of the Ig heavy- and light-chain genes,
key players of innate immunity in their role as opsonizers serving to greatly increase the affinity of antibody mole-
that enhance phagocytosis and as effectors of ADCC and cules for their cognate antigens [see below].)
of the classical pathway of complement activation. Thus,
it is expected that viruses would have evolved counter-
measures that impair these antibody-dependent innate THE ADAPTIVE IMMUNE SYSTEM
immunity defenses by targeting the antibody. HSV and Macrophages, neutrophils, NK cells, IFN, and comple-
other herpesviruses provide an example of a viral coun- ment of the innate immune system slow down most virus
termeasure that targets the antibody. As noted above, infections and may even be sufficient to contain some of
these viruses produce envelope glycoproteins that bind them. However, in most cases, virus infections progress to
the Fc portion of antibody molecules, by that means pro- a point where the innate immune system needs reinforce-
tecting HSV from opsonin-enhanced phagocytosis. More- ments. Under those circumstances, the adaptive immune
over, those viral glycoproteins protect HSV-infected cells response is activated.
from ADCC and also protect both free virus particles and In point of fact, because all viruses exploit an intracel-
infected cells from complement-mediated lysis via the lular lifestyle, the innate immune system is generally much
classical complement pathway. less effective at overcoming viruses than it is at overcom-
Viruses have other strategies for evading the effects of ing pathogens that are strictly extracellular. Fortunately
antibodies that are more properly considered in the con- for us, effectors of the adaptive immune system are espe-
text of evading adaptive immunity. Some of these are sim- cially well suited to the task of clearing virus infections.
ply mentioned here and discussed in more detail later. The killer cells or CTLs of adaptive immunity are special-
They include (i) antigenic variation via several stratagems; ized to attack and destroy host cells that harbor intracel-
(ii) being inherently poor antigens; (iii) replicating in im- lular virus, while B cells produce soluble factors (antibod-
munoprivileged sites (e.g., the central nervous system); ies) that specifically neutralize extracellular virus. Together,
and (iv) establishing latent infection, in which virus pro- the T cells and B cells of adaptive immunity work to inca-
teins are not expressed. pacitate both intracellular and extracellular virus. More-
over, the adaptive immune system has a memory of
APOBEC3G and the HIV Vif Protein its encounter with each pathogen. This immunological
The HIV Vif (virion infectivity factor) protein is neces- memory allows the adaptive immune system to respond
sary for HIV to replicate in human T lymphocytes. Recent faster and stronger to a subsequent encounter with the
104 CHAP TER 4

same pathogen, a phenomenon referred to as acquired antiviral antibodies alone are generally not sufficient to
immunity, which may be lifelong against some viruses. eliminate viral pathogens.
Adaptive immunity also can be important in overcom- Fortunately, intracellular virus can be eliminated through
ing bacterial infections. Antibodies of adaptive immunity the action of the second branch of adaptive immunity, the
act against all bacteria, and CTL-mediated defenses may cell-mediated immune response. CTLs, which mature in
be important against the relatively few bacteria that have the thymus, are the principal effectors of cell-mediated
an intracellular lifestyle (e.g., the chlamydiae). And, as immunity against viruses. These cells specifically recog-
might be expected, intracellular bacteria, like viruses, have nize and destroy virus-infected cells, destroying nascent
evolved countermeasures against CTL-mediated defenses. intracellular virus along with the cell. The sacrifice of an
One such countermeasure used by intracellular bacteria is infected cell is more than compensated for by preventing
to reside in an intracellular compartment that they mod- the spread of the infection to healthy cells.
ify so that it does not intersect the pathways of antigen Generally speaking, while the cellular immune response
presentation by MHC molecules (see below). Yet bear in is generally necessary to resolve a viral infection, the hu-
mind that all viruses are obligate intracellular parasites, moral and cellular arms of adaptive immunity in fact co-
and consequently, the objective of the adaptive immune operate in that effort. In addition, having the two arms
response to a virus infection is to eradicate virus-infected provides a safety net, should one of the arms be impaired.
cells as well as free extracellular virus particles. Because it Nevertheless, the relative importance of humoral versus
is crucial to eliminate intracellular as well as extracellular cellular immunity actually varies from one virus to an-
virus, and because the adaptive immune system is uniquely other. This is readily demonstrated using knockout mice,
suited to destroying intracellular virus, it is plausible that in which different components of the immune system are
the adaptive immune response may have evolved primar- specifically depleted. In addition, humans with agamma-
ily to respond to the threats posed by viruses. globulinemia (i.e., impaired synthesis of Igs) provide a
The adaptive immune response is so named because it real-life model for assessing the importance of humoral
adapts to each specific pathogen that it encounters, tailor- immune responses against particular virus infections. For
ing a specific response to each. As noted above, adaptive instance, these individuals are more at risk for developing
immunity is mediated by B cells and T cells, which con- paralytic poliomyelitis after receiving the live oral polio
tain cell surface receptors that are specific for particular vaccine (Chapter 6). Moreover, patients with agamma-
antigens. Moreover, these receptors vary from one T cell globulinemia may continually excrete enteroviruses, in-
to another. This is in contrast to NK cells of innate im- cluding vaccine-derived poliovirus, for extended times.
munity, which contain a fixed set of receptors that are the (There is some concern that long-term poliovirus shed-
same on all NK cells. ders, although rare, might in the future create a risk to the
Adaptive immunity comprises two distinct but com- general population at a time after poliovirus might be
plementary branches for containing and eliminating an considered to have been eradicated and routine vaccination
infection. One branch of adaptive immunity, referred to is no longer practiced.)
as humoral immunity, is based on antibodies, which are A most important feature of the adaptive immune re-
synthesized by B cells (i.e., lymphocytes that mature in the sponse noted above is that it has a memory. That is, it re-
bone marrow). The other branch of adaptive immunity, members its first or primary encounter with a pathogen.
the cell-mediated immune response, is carried out by Because of this memory feature, the adaptive immune
T cells (i.e., lymphocytes that mature in the thymus). system can quickly mount a vigorous immune response to
The mechanisms by which antibodies control virus in- a previously encountered pathogen, in just a matter of
fections are described below. For now, a major role for hours after reexposure. This aspect of the memory re-
antibodies is to directly bind to viruses, thereby prevent- sponse is in contrast to the primary response, which can
ing them from infecting susceptible cells. An antibody take many days to reach effective levels. The rapid-recall
that does that is said to neutralize the virus. memory response can prevent or greatly diminish clinical
A shortcoming of neutralizing antibodies is that the symptoms. It is the basis for acquired immunity and also
infection inside a cell is generally inaccessible to antibod- the rationale underlying vaccination. Specificity and mem-
ies. Thus, intracellular virus replicates in an antibody-free ory are the two hallmarks of adaptive immunity.
environment, generating thousands of progeny virus par-
ticles that are released into the extracellular environment. Antibodies and B Cells
A sufficient number of these released viruses may escape Antibody responses against viruses are generally detect-
being neutralized by extracellular antibodies and succeed able by 3 to 5days after infection, and antivirus antibodies
in beginning new replication cycles in new cells. Thus, have been shown to help clear primary infections involving
Host Defenses and Viral Countermeasures 105

a variety of different viruses. Moreover, they are a key de- sites. However, this premise is now controversial because
fense against previously encountered viruses, and in the it could not be confirmed by structural studies.
cases of viruses that establish latent or persistent infec- Second, antibody molecules have multiple antigen-
tions, antibodies can control new outbreaks in an indi- binding sites (Fig. 4.4). Consequently, antibodies can ag-
vidual that may occur after a period of viral inactivity. The glutinate virus particles. This reduces the number of in-
ability of antibodies to help contain persistent virus infec- fectious units and facilitates clearance of the virus by
tions is largely due to immunological memory. phagocytes.
Antibodies can impede a virus infection by several dif- Each of the antibody mechanisms enumerated above is
ferent mechanisms (Fig. 4.8). First, some antibodies may said to neutralize the virus, that is, impair its infectivity.
interfere with the ability of a virus to attach to, and infect, Recall that antibodies also may interact with effectors of
a host cell. This can happen if the antibody binds directly innate immunity to help contain the infection. For exam-
to the viral attachment sites. For some time it was believed ple, antibodies may opsonize the virus and participate in
that other antibodies, which bind elsewhere on the viruss ADCC and complement fixation.
surface, may yet block infection by causing conformational Despite the notion that antibodies act against extracel-
changes on the virus particle that affect its attachment lular viruses only, antibodies can disrupt the intracellular
replication cycles of some viruses. For example, antibod-
ies against measles virus can suppress measles virus tran-
scription and translation, at least in cell culture. The un-
Figure4.8 Antibodies can impede a virus infection by several differ-
ent mechanisms. (A) Some antibodies may interfere with the attach- derlying mechanism is not clear, but it may be the result of
ment of a virus to the host cell. This can occur if the antibody binds an effect mediated by the antibodies from the cell surface.
directly to the viral attachment sites. (B) It once was believed that In another example, passive transfer of antibodies against
some antibodies may cause conformational changes on the virus Sindbis virus can clear that virus from the brains and spi-
particle that affect but do not directly block its attachment sites. This
premise is now controversial because it could not be confirmed by nal cords of persistently infected SCID mice (see Box4.4
structural studies. (C) Aggregation of viruses by antibodies reduces for a description of SCID mice). The mechanism by which
the number of infectious units and facilitates clearance of the virus antibodies clear the Sindbis virus infection is not entirely
by phagocytes. (D) Contrary to the notion that antibodies act only known, but studies invitro show that it requires antibody-
against extracellular viruses, in some instances antibodies appear to be
able to disrupt the intracellular replication of viruses, by mechanisms mediated cross-linking of viral glycoproteins on the cell
that are not entirely clear. (E) Opsonization. (F) Lysis of virus/anti- surface and involves inhibition of virus budding (Bur-
body complexes by complement. (G) Lysis of virus-infected cells by deinick-Kerr etal., 2007). Thus, these antibodies against
antibody plus complement. (H) Antibody-dependent cellular cytotox- Sindbis virus impair the maturation and release of intra
icity (ADCC). (I) Inhibition of virus budding and release. Abbrevia-
tions: VR, virus receptor; V-ag, viral antigen; mj, macrophage; CR2, cellular Sindbis virus.
complement receptor. Adapted from H. E. Kaufmann, A. Sher, and All antibody molecules are comprised of four polypep-
R. Ahmed (ed.), Immunology of Infectious Diseases (ASM Press, tide chains: two identical heavy chains and two identical
Washington, DC, 2002), with permission. smaller light chains, which are linked together by disulfide
bonds (Fig. 4.4). This combination of heavy and light
Virus chains appears as a Y-shaped structure. Each branch of the
C Y is comprised of a complete light chain and the N-terminal
Antibody portion of a heavy chain. The stem of the Y contains the
Complement VR
paired C termini of the heavy chains.
Lysis Before going on with our description of antibody mol-
CR2 ecules, it would be useful to specify more precisely what
F antibodies actually recognize. Up to now, we have been
using the term antigen without rigorously defining it.
FcR Actually, antigen is a rather loose term referring to an an-
E tibody target. At one extreme, an antigen may be the whole
V-ag pathogen, or it may be a particular protein. Importantly,
I H each antibody (and T-cell receptor as well; see below) rec-
ognizes and binds to a specific site or determinant on the
antigen, referred to as an antigenic determinant or epitope.
NK The epitope recognized by an antibody is comprised of a
cluster of only 3 to 5 amino acids. (In contrast, T-cell re-
ceptors recognize between 8 and 20 amino acids.) Thus,
106 CHAP TER 4

even a small protein may contain numerous epitopes, and Like the variable domains, the constant (C) domains,
as a result, a protein can be recognized by several different which make up the stem of the antibody molecule, also
antibodies, each of which is specific for a particular are comprised of Ig domains. However, in contrast to the
epitope on the protein. variable domains, there is little or no sequence diversity in
The antigen-binding sites on an antibody molecule are the constant domains within an antibody class.
located at the ends of the branches of the Y-shaped struc- The sequences within a V domain that actually deter-
ture. Thus, each antibody molecule has two antigen-binding mine epitope-binding specificity are comprised of three
sites. Also, amino acids on both a heavy chain and a light loops at the end of the V domain referred to as hypervari-
chain contribute to each antigen-binding site. As we soon able regions, and also as complementarity-determining
will see, each of the antigen-binding sites on an antibody regions (CDRs), because they have a surface that is com-
is specific for the same epitope. Consequently, antibodies plementary to their cognate epitopes (Fig. 4.10). The
are able to aggregate viruses into clusters, thereby reduc- CDRs are flanked by so-called framework regions.
ing the number of infectious units and facilitating clear-
ance of the virus by phagocytes. Moreover, when antibod- Antibody Diversity
ies bind to their cognate antigens, the single invariant Here is an interesting question. Since each antibody mol-
portion, referred to as the Fc region (see below), is avail- ecule is specific for a particular epitope, how many differ-
able to interact with other immune effectors. Recall that ent antibody specificities are available within an individ-
neutrophils, macrophages, and NK cells all have receptors ual? Well, current estimates range from as high as 108 to
on their surfaces that can bind to the free Fc regions of an astonishing 1016 different specificities! Presumably, this
antigen-bound antibody molecules. Furthermore, bound variety of antibody specificities available within an indi-
antibodies can trigger the classical complement pathway. vidual, known as the antibody repertoire, suffices to rec-
There are five functionally distinct classes of antibod- ognize any pathogen that might be encountered.
ies. They are IgA, IgD, IgE, IgG, and IgM. These antibody This brings up what may be a more compelling ques-
classes are distinguished at the molecular level by differ- tion. Considering that every antibody is a protein, how
ences in their stem structures. Consequently, the particular many genes might be required to generate 108 different an-
specialized functions of these different antibody classes are tibodies (the conservative estimate of our antibody reper-
determined by their stem structures, as discussed below. toire)? Well, it is much less than 108, which is fortunate for
IgG antibodies, which are the most common antibody us since we only have about 3 105 genes. But recall that
class in the blood, are generally used to illustrate common each antibody variable region is comprised of a variable do-
structural features of antibodies (Fig. 4.4). IgG antibodies main from a heavy chain (VH) and a variable region from a
contain a relatively unstructured tract in the middle of the light chain (VL). Thus, mixing and matching 104 heavy
heavy chain, which forms a flexible hinge region. This is chains with 104 light chains could do the trick. But that
useful for our description of antibodies because the pro- would still require a lot of genes (i.e., 2 104, or about 10%
tease papain cleaves IgG molecules at the hinge region to of our entire genome). Fortunately, elegant mechanisms
generate three fragments: two Fab fragments and one have evolved which generate a seemingly infinite number
Fc fragment. The Fab (acronym for fragment antigen- of antibody specificities from many fewer genes than that!
binding) fragments contain the antigen-binding sites on The immune system uses two general principles, which
the antibody, whereas the Fc (acronym for fragment- it builds upon, to generate this virtually limitless number
crystallizable, because it readily crystallized in early ex- of different antibody specificities. The first of these prin-
periments) fragment provides the link to other immune ciples is modular design, and the second principle is
effectors. (It may be useful to think of the Fc fragment as clonal selection. Modular design accounts for how a pop-
the constant fragment, since it varies little, if at all, within ulation of nave B cells (i.e., B cells that have never en-
an antibody class [see below].) countered an antigen) preexists in the host, ready with
The heavy and light chains each contain reiterations of receptors that can recognize and respond to virtually any
a particular sequence motif referred to as an Ig domain new invader. Clonal selection explains how when nave B
(Fig. 4.9). Each such domain is about 100 amino acids in cells encounter their cognate antigen, they expand into a
length. Notice how Ig domains consist of two -sheets clone of identical B cells that secrete antibody molecules,
that are connected by loops. An Ig domain that partici- all with the same specificity.
pates in antigen binding is referred to as a variable (V) We begin by considering modular design, i.e., the
domain. V domains are so named because they contain mechanism for generating a population of nave B cells,
the sequence differences between antibody molecules that each member of which is specific for a particular epitope,
enable them to specifically recognize their cognate epitopes. such that the B-cell population as a whole may have
Host Defenses and Viral Countermeasures 107

N terminus


C terminus
strand strand

Disulfide Disulfide
bond bond

Light-chain C domain Light-chain V domain

Figure4.9 The heavy and light chains of an Ig molecule contain four and two Ig domains, respectively.
Shown is an individual light chain of an IgG molecule. Each Ig domain contains about 100 amino acids,
which are folded into two sheets that are connected by loops. An Ig domain that participates in antigen
binding is referred to as a variable (V) domain (right panel). V domains are so named because they con-
tain the sequence differences between antibody molecules that enable them to specifically recognize their
cognate epitopes. Like the variable domains, the constant (C) domains, which make up the stem of the
antibody molecule, also are comprised of Ig domains. However, in contrast to the variable domains, there
is little or no sequence diversity in the constant domains within an antibody class. Arrows point from the
amino terminus to the carboxy terminus. The inset depicts the location of the light chain in an IgG mol-
ecule. Adapted from P. Parham, The Immune System, 2nd ed. (Garland Science, New York, NY, 2005),
with permission.

members ready to respond to virtually any pathogen that heavy-chain gene by splicing together a V, a D, a J, and a C
might be encountered. The underlying basis for modular segment (Fig. 4.12). The combined V, D, and J segments
design is that the Ig genes are arranged on the chromo- encode the variable region, and the C segment encodes the
some in a special kind of fragmented form, from which constant region. And, a mature light-chain gene is assem-
mature Ig genes may be assembled by a special genetic re- bled from a V segment and a J segment.
combination process that occurs only in B cells (Fig. 4.11 Now here is a really important point. Many alterna-
and 4.12). More precisely, a B cell assembles a mature tive forms of each of the V, D, and J segments are present

Antigen- Figure4.10 The hypervariable regions (HV1,

binding HV2, and HV3) of an antibody molecule, which
site determine epitope-binding specificity, are com-
prised of three loops at the end of the V domain
N that also are referred to as complementarity-
HV3 (CDR3) determining regions (CDRs), because they have
a surface that is complementary to their cognate
epitopes. These regions are flanked by so-called
framework regions. Adapted from P. Parham,
HV1 (CDR1) The Immune System, 2nd ed. (Garland Science,
New York, NY, 2005), with permission.

HV2 (CDR2)
108 CHAP TER 4

light-chain locus
L1 V1 L2 V2 L V~ 30 J1 C 1 J2 C 2 J4 C 4

Chromosome 22

light-chain locus
L1 V1 L2 V2 L3 V3 L V~ 40 J15 C

Chromosome 2

Heavy-chain locus
LH1 VH1 L2 VH2 L3 VH3 L VH ~ 65 DH127 JH16 C

Chromosome 14
Figure4.11 The basis for modular design is that the Ig genes are arranged in the genome in a special kind of frag-
mented form. In the human genome there are two loci, and (kappa), that encode light-chain segments, and one
locus that encodes heavy-chain segments. Different Ig gene segments encode the leader peptide (L), the V region (V),
and the constant region (C) of both the heavy and light chains. Note the multiple number of different V gene seg-
ments for the light and heavy chains. Also note the multiple joining (J) gene segments and the additional set of diver-
sity (D) gene segments in the heavy-chain locus. The heavy-chain locus also contains C gene segments for each of the
antibody classes. Introns separate each of these gene segments. A mature light-chain gene contains one V segment and
one J segment, which together constitute the V region, and one C segment. A mature heavy-chain gene is comprised
of a V, a D, a J, and a C segment. These are joined together by DNA recombination (Fig. 4.12). Humans have two func-
tionally equivalent light chain isotypes, kappa (k) and lambda (l), and five functionally distinct heavy chain classes (see
text). Adapted from P. Parham, The Immune System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.

within the unrecombined DNA (Fig. 4.11). Moreover, One of the 27 D segments is recombined with one of the
these alternative forms can be spliced together by DNA 6 J segments. One of the 65V segments is then joined to
recombination such that each of the alternative V, D, and the DJ sequence, and the resulting VDJ sequence is then
J segments is selected at random. The course of events joined to a C segment (Fig. 4.12). The mature light gene
for generating the mature heavy-chain gene is as follows. is likewise assembled using this mix-and-match strategy.

Figure4.12 Many alternative forms of each of the V, D, and J segments are present within the germ line DNA
(Fig. 4.11). Importantly, these alternative forms can be spliced together by DNA recombination such that one of each
of the alternative V, D, and J segments is selected at random. The course of events for assembling a mature light-chain
gene is as follows. One of 30 to 40V segments is joined to one of the J segments to form a complete V region, which is
then joined to a C segment. The course of events for generating the mature heavy-chain gene is as follows. One of the
27 D segments is recombined to one of the 6 J segments. One of the 65V segments is then joined to the DJ sequence,
and the resulting VDJ sequence is then joined to a C segment (Fig. 4.10). Adapted from P. Parham, The Immune
System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.
Light chain Heavy chain
Germline DNA Germline DNA

Somatic DJ-joined DNA

VJ-joined VDJ-joined
rearranged DNA rearranged DNA
Host Defenses and Viral Countermeasures 109

These rearrangements of heavy- and light-chain gene loci 2,000molecules per s) of the soluble form of the antibody
occur only in B cells, and the process is referred to as somatic into the blood and tissues. Activation of the B cell and its
recombination, since it does not occur in the germ line. clonal expansion may take a week or longer. During this
Because of the number of different V, D, and J DNA period, the innate immune system is responsible for con-
segments that are recombined together at random to make taining the infection.
up a variable region, this strategy generates about 107 We noted above that Mother Nature has evolved a way
different antibody specificities, which is an impressive to improve upon modular design, to actually perfect anti-
number, but not quite enough to account for the estimated bodies over the course of an infection. This happens as
number of different antibody specificities. Well, additional follows. When a nave B cell encounters its cognate anti-
diversity is created as a consequence of the imprecision of gen and is activated to undergo clonal expansion, its
the recombination process. Because of this imprecision, ex- V-domain coding sequences go through a process called
tra nucleotides are inserted at the recombinational joints. somatic hypermutation, which randomly introduces
Thus, evolution has given rise to a means for generating a point mutations at a high rate (as high as one mutated
seemingly limitless number of antibody specificities from base per 1,000 bases per generation) throughout the re-
only about 300 genes! However, that is still not the end of arranged heavy- and light-chain genes. Importantly, so-
the story, since Mother Nature has invented a way to im- matic hypermutation is restricted to the variable regions
prove the affinities of those antibodies for their cognate of the rearranged Ig genes. The process involves an en-
antigens over the course of an infection, as will be seen zyme, activation-induced cytidine deaminase, which is
shortly. Moreover, recalling that there are five functionally induced when the nave B cell is activated. Activation-
distinct classes of antibodies, each of which is specialized induced cytidine deaminase converts cytidine residues
for a particular purpose, the immune system selects the in the DNA to uracil. The uracil residues are excised, and
antibody class that is most appropriate for the particular replacement nucleotides are randomly incorporated
circumstance. when the DNA is replicated. (Recall that T cells express
An important consequence of modular design is that a cytidine deaminase, APOBEC3G, which can hyper
each B cell makes identical antibodies that are specific for mutate the HIV genome, rendering it noninfectious [see
one particular epitope. That is, each B cell makes heavy above].)
chains with only one VH region and light chains with only Somatic hypermutation generally leads to antibody V
one VL region. You may have noticed a bit of a conundrum regions that have higher affinities for their cognate epitopes
here because of the fact that B cells are diploid, with mod- than those encoded by the original rearranged Ig genes.
ular rearrangements likely occurring independently at This happens as follows. Recall that a transmembrane
each of the heavy- and light-chain loci. Modular rear- form of the antibody is present as the B-cell receptor. Those
rangements indeed do occur independently at each heavy- B cells that express higher-affinity receptors are selectively
and light-chain locus. However, a successful rearrange- activated for further clonal expansion. This occurs because
ment occurs first at one of the loci. When that happens, maturing B cells must continually be restimulated by
the other locus is silenced. binding to their cognate antigen in order to continually
Clonal selection, the second general principle underly- proliferate. Consequently, those B cells that express the
ing the generation of antibody-producing B cells, occurs highest-affinity B-cell receptors are preferentially selected
as follows. When a B cell matures and begins to make its to continue dividing and become antibody-producing
particular Ig, the heavy chains contain a hydrophobic se- plasma B cells. The end result is that the antibody response
quence near their carboxy termini, causing the antibody is fine-tuned to fight the invader. The process is called af-
molecules to become integral membrane proteins that finity maturation. This is still not the end of the story.
are displayed on the surface of the B cell. No soluble anti- Recall that there are five different classes of antibodies:
bodies are yet released by these cells. However, the surface IgA, IgD, IgE, IgG, and IgM. These classes are distinguished
antibodies, called B-cell receptors, are able to recognize by their different constant regions. Importantly, whereas
their cognate antigen. When the B-cell receptor of a par- the variable regions determine an antibodys specificity,
ticular nave B cell (i.e., a mature B cell that has left the the constant regions specify how and where the antibody
bone marrow but has not yet encountered its cognate an- will function. This plays out as follows.
tigen) encounters and binds its cognate antigen, a signal is When a B cell is born in the bone marrow, it rearranges
transmitted into the B cell, causing it to proliferate and its V, D, and J Ig gene segments and initially joins them to
generate a clone of about 20,000 identical B cells. Most of an M-type constant region. Consequently, the B-cell re-
the B cells in the clone mature into plasma B cells, each ceptors are initially of the IgM class, and when the B cell is
of which produces and secretes huge quantities (about activated it will initially produce only IgM antibodies.
110 CHAP TER 4

This is advantageous to the host, because IgM antibodies The same rearranged Ig gene encodes IgM and IgD an-
are exceptionally good at fixing complement, which may tibodies, but the messenger RNA (mRNA) is spliced one
be important at a time when antibodies are just beginning way to encode an IgM antibody and another way to en-
to be made. In addition, IgM antibodies are generally code an IgD antibody. IgD antibodies make up only a
good at neutralizing viruses, although affinity maturation small percentage of the circulating antibodies in humans
is sometimes necessary. Importantly, however, the B cell and are found in small amounts in the tissues that line the
has the opportunity to change the class of antibody it belly and chest. How they might work is not clear. We have
makes from IgM to any of the other classes, each of which little to say here about IgE antibodies, since they confer
provides special advantages. protection against parasites (e.g., worms).
IgG antibodies also are very good at neutralizing vi- As noted above, B cells initially produce IgM antibod-
ruses and offer the additional unique asset of being able to ies. Class or isotype switching, to produce other antibody
pass from the mothers blood across the placenta into the classes, is accomplished at the genome level by splicing an
fetal circulation. These maternal antibodies provide the IgA, IgG, or IgE C gene segment to the VDJ sequence that
newborn with a measure of protection against currently was assembled for the original IgM antibody. The original
circulating pathogens, at a time when the newborns own IgM C segment is spliced out in the process. Notice that by
immune system needs to become operational. The fact means of affinity maturation and class switching, the B cell
that IgG antibodies are much longer-lived than IgM anti- is able not only to fine-tune the specificity of the antibody
bodies also suits them to this purpose. Moreover, NK cell but also to fine-tune it with respect to its sites of action and
receptors specifically interact with the Fc segment of IgG function. But that is still not the end of the story.
antibodies. Consequently, this is the antibody class that As the clone of activated B cells proliferates and ma-
mediates ADCC. tures, not all of its members become antibody-producing
Whereas IgG antibodies are the most common anti- plasma cells. Instead, some of those B cells mature into
bodies in the blood, IgA antibodies are the most abun- memory B cells. Importantly, upon subsequently reen-
dant antibody class in the body. The reason for the abun- countering the same pathogen, these memory cells elicit a
dance of IgA antibodies is that these antibodies guard the much faster and stronger immune response. There is ac-
mucosal surfaces of the body and the human body has tually more to the story of humoral immunity than these
about 400m2 of these mucosal surfaces to defend. There memory B cells. In addition, there is a population of
are more IgA-producing B cells located under the mucosal plasma B cells that persist, constitutively producing anti-
epithelia than there are B cells in any other organ of the bodies (see below). The preexisting antibodies produced
body, including the spleen, thymus, and lymph nodes, by these cells provide the first line of defense against rein-
which are the major sites in the body where cells of the fection with a previously encountered microbial patho-
immune system interact. Bear in mind that these delicate gen. (T cells likewise generate a population of memory
mucosal surfaces, which include those of the respiratory, cells with attributes similar to those of memory B cells
digestive, and urogenital tracts, are all that separate those [see below].)
tracts from the outside world. Since the rest of the body
surface is a mere 2m2 and is covered by dead keratinized Viral Evasion of Antibodies
skin cells, the mucosal surfaces are where the vast major- So, we have a preexisting population of nave B cells that
ity of viruses invade the body. Thus, one can appreciate is diverse enough to recognize and respond to virtually
why they are so immunologically active. IgA antibodies any microbial invader. Moreover, we have a population of
also are secreted into the milk of nursing mothers, thereby memory B cells ready to spring into action more quickly
protecting the babys digestive tract against currently cir- and strongly should we reencounter an earlier invader. In
culating pathogens that might be ingested. addition, after a primary infection by a pathogen, a popu-
IgA molecules reach their site of action on the mucosal lation of plasma cells remains that constitutively produces
surface by a process known as transcytosis, whereby they antibodies against that pathogen. How then do some vi-
first bind to an IgA receptor on the basolateral surface of ruses still manage to evade humoral immunity and cause
an epithelial cell and then undergo receptor-mediated illness, and why, in some instances, do we have multiple
transport to the apical side of the cell, where they are re- episodes with the same virus?
leased. Importantly, IgA antibodies do not fix comple- Regarding the illness issue, in some instances the im-
ment. This is good, considering all the foreign microor- mune response itself contributes significantly to the symp-
ganisms coming into contact with mucosal surfaces. Thus, toms of the infection (see below). Yet there also are a vari-
if IgA antibodies fixed complement, then mucosae would ety of different strategies that viruses have evolved to
be in a constant state of inflammation. evade humoral immunity. For example, we already noted
Host Defenses and Viral Countermeasures 111

how herpesviruses express receptors for the Fc domains of Humoral immunity likely plays an important part in
antibodies, enabling them to thwart ADCC and anti- maintaining the long period of clinical latency, since huge
body-mediated interactions with complement. Some numbers of virus particles are neutralized, most probably
additional examples of viral antibody evasion strategies by antibody-mediated opsonization and phagocytosis. Yet
are as follows. the high mutation rate of HIV ensures that an untreated
A major means by which viruses evade antibody- patient will have viruses circulating that contain virtually
mediated host defenses is by generating so-called immune every possible base substitution. Among these will be a
escape mutants, which display altered antigenicity. In ac- large variety of immune escape mutants.
tuality, a high mutation rate is virtually an inevitable fea- The high HIV mutation rate is but one of several means
ture of the viral lifestyle. In the case of RNA viruses, their this virus has for evading the humoral immune response
RNA-dependent RNA polymerases do not have a proof- and ultimately destroying adaptive immunity. The HIV
reading function (as DNA polymerases do) in the form of envelope glycoprotein gp120, which is the predominant
an exonuclease that can excise misincorporated nucle- antigen on the viral surface, has features that make it a
otides. Consequently, their error frequency can be as high difficult target for neutralizing antibodies from the very
as one misincorporated nucleotide for every 103 to 104 nu- start of the infection. This protein is one of the most ex-
cleotides incorporated into RNA. Thus, essentially every tensively glycosylated proteins known, making it a weak
individual virus in an RNA virus population has a unique antigen. That is, relatively little neutralizing antibody is
RNA genome sequence. DNA viruses are more genetically generated during an HIV infection. HIV has several other
stable than RNA viruses, but even they display mutation immune escape mechanisms that target cell-mediated im-
rates several orders of magnitude greater than that of the munity (see below).
host. Making matters worse yet for the host, all viruses Human rhinoviruses, which cause the common cold
generate huge numbers of progeny in a relatively short pe- (Chapter 6), use several different strategies to evade neu-
riod of time. Bearing in mind that a change in only one tralization by antibodies. First, the rhinovirus receptor-
amino acid at a neutralizing antibody-binding site can binding sites, on the surface of their icosahedral capsids,
drastically reduce the affinity of an antibody against that are recessed in what are referred to as canyons. Impor-
site, we can appreciate how immune escape mutants can tantly, these canyons are narrower than the footprint
readily emerge. Thus, the humoral immune response is of- made by the antigen-binding site of an antibody. Thus,
ten playing a game of catch-up, continually having to am- the receptor-binding sites, while accessible to the receptor,
plify its response against each new serological variant of the are inaccessible to antibodies. Consequently, these viruses
virus. Generally, immune escape mutants may be more of a are somewhat resistant to potentially neutralizing anti-
problem in the cases of RNA viruses than for DNA viruses, bodies that might recognize those sites. Moreover, having
because of the higher mutation rates of the former. But im- receptor binding sites that are inaccessible to antibodies
mune escape mutants are an issue in the cases of all viruses, implies that these viruses are under little, if any, selective
DNA as well as RNA, that establish persistent infections. pressure from the host immune system to modify their
HIV is especially adept at evading nearly all measures receptor binding sites, something they could not likely do
of host immunity, employing a number of different strat- and remain viable.
egies toward that end. The generation of immune escape Second, there are more than 100 antigenically distinct
mutants is one of the key means by which that virus evades rhinovirus serotypes circulating in humans, so that there
antibody-mediated defenses. HIV infections begin like are always some serotypes against which a person may
many other virus infections, with an early acute phase have little, if any, acquired immunity. Third, some sources
characterized by extensive virus replication. This acute state that the innate immune system is often able to re-
phase is followed by a marked decrease in virus levels, as solve rhinovirus infections within just a few days, so that
the adaptive immune response comes into play to bring the adaptive immune system may not get fully activated. If
the infection under control. However, HIV persists in in- that is so, then these viruses yet remain quite successful
fected individuals for years and is rarely and perhaps never since they are so efficiently transmitted that the infection
eliminated. That is so, despite the fact that infected indi- can be passed on, even in the short time before it is cleared
viduals mount vigorous adaptive immune responses against by innate immunity. In fact, the viruses may benefit, since
the virus until nearly the end. During much of this time, if the adaptive immune system is not activated, then there
the patient does not display symptoms. This period of clin- is no acquired immunity, and one can be reinfected with
ical latency can last for 10years or more. However, again the same rhinovirus serotype. (Yet I believe it is unlikely
note that this period is actually characterized by immense that any virus infection might be resolved completely in
levels of immune activity, as well as virus replication. the absence of adaptive immunity.)
112 CHAP TER 4

In contrast to the rhinovirus case, in which more than

100stable serotypes may circulate simultaneously in the
Box 4.8
human population, there are usually only one or two pre-
The term immune privilege is meant to convey the idea
dominant strains of influenza virus that circulate in hu-
that there is an absence of immunosurveillance within
mans at any time (Chapter 12). Yet influenza virus under- certain organs and tissues, such as in the central nervous
goes continuous antigenic change from one year to the system. This notion is now known to be too extreme.
next, such that acquired immunity does not carry over For example, tissue grafts within the brain are eventually
well from year to year, thereby ensuring the survival of the rejected. Also, there is the important example of experi-
virus in the population. There are two basic means by mental autoimmune encephalitis, in which immunizing
which influenza virus undergoes antigenic change to keep mice with myelin basic protein (a principal protein of
ahead of acquired immunity. The first is the simple accu- the myelin sheaths in the brain) induces brain invasion
mulation of mutations that results from the high rate of by lymphocytes, destruction of myelin, and paralysis.
misincorporation of nucleotides while replicating the Experimental autoimmune encephalitis is considered to
influenza virus RNA genome. This type of change is re- be an important model of multiple sclerosis (see the text
ferred to as antigenic drift. The second type of change, and Box4.19). Yet autoimmune attack on our myelin is
referred to as antigenic shift, depends on the fact that the fortunately rare. What factors might account for that?
influenza virus genome is comprised of eight distinct The answer appears in part in the text.
RNA segments. For this reason, if a host were simultane-
ously infected with two different influenza virus strains,
their gene segments might reassort with very high fre-
quency to yield a completely new strain. Serious human not express MHC molecules, which are needed to activate
pandemics (worldwide epidemics) are thought to result and mediate adaptive immunity (see below). (Note that
in part from the genetic reassortment of human influenza dendritic cells, like macrophages, are sentry cells sta-
virus strains with avian influenza virus strains. Humans tioned beneath the mucosa and other epithelia. Although
may have little to no immunity to these reassortant vi- they have many properties in common with macrophages,
ruses. (This issue is discussed at length in Chapter 12.) an important difference is that while activated mac-
Yet another strategy for evading humoral immune re- rophages remain in place at an infected site, activated den-
sponses is to replicate within cells or tissues that are said dritic cells migrate from the infected site to the nearest
to be in immunologically privileged sites, that is, sites draining lymph node, where they activate B cells and T
that are subject to reduced immune surveillance. The kid- cells of adaptive immunity.)
ney tubules are such a site. They are continually exposed Fortunately for us, tight junctions between the cells
to foreign matter, and they employ measures other than that line the brain capillaries and ventricles constitute a
immune surveillance to avoid a continuous state of in- so-called blood-brain barrier, which limits the entry of
flammation. Possibly, it is for this reason that the kidneys many pathogens from the blood circulation into the brain.
are a favored site for the replication of several polyomavi- Yet the blood-brain barrier does allow the limited entry of
ruses, including simian virus 40, JC virus, and BK virus. immune cells circulating in the periphery (Box4.8). Also,
Reduced immune surveillance of the kidney tubules also resident glial cells in the brain can function as antigen-
may be a factor that favors the establishment of persistent presenting cells, the significance of which is explained later.
infections by these polyomaviruses. Nevertheless, and predictably perhaps, the factors that
As explained below, activation of the adaptive immune make the brain a somewhat immunologically privileged
response generally requires that antigen or antigen- site together make it a favored site for some viruses that
presenting cells enter lymph nodes, where antigen can be establish persistent infections there. The herpesviruses,
presented to helper T cells. Importantly, the brain does not polyomaviruses, complex retroviruses (Chapter 21), and
have classical antigen-presenting cells such as dendritic papillomaviruses are examples of virus families that thrive
cells, and it does not drain through conventional lym- by establishing persistent infections, often in immuno-
phatics (see the note at the end of this paragraph). Thus, logically privileged sites, such as the central nervous
infection of brain sites does not readily lead to activation system. HIV and HSV are important examples of viruses
of adaptive immunity (Box 4.8). Moreover, it is reason- that establish persistent infections in the central nervous
able to suggest that the brain is devoid of mediators of system. Moreover, the ability of some of these viruses to
inflammation, since inflammation within the brain could establish a latent state in some cells, in which viral genes
be disastrous for the host. Another important factor mak- are temporarily not expressed, contributes to their ability
ing the brain an immunoprivileged site is that neurons do to evade immune detection and their success invivo.
Host Defenses and Viral Countermeasures 113

The neurotropic viruses enumerated above are able to MHC molecules is one of the central and perhaps most el-
gain access to the central nervous system despite the blood- egant aspects of the adaptive immune response. (Impor-
brain barrier, doing so by several different routes. For ex- tantly, in contrast to macrophages and neutrophils of the
ample, viruses in the blood circulation might infect and innate immune system, which recognize regularly repeat-
disrupt the capillary endothelial cells or the meninges that ing arrays of proteins and carbohydrates that are shared
constitute the blood-brain barrier. HSV, varicella-zoster among broad classes of microbes, lymphocytes of the adap-
virus (a herpesvirus responsible for chicken pox and shin- tive immune system specifically recognize small segments
gles), and rabies virus employ another stratagem to access of proteins, either directly as in the case of B cells or follow-
the central nervous system. They initially infect mucoepi- ing processing and presentation as in the case of T cells.)
thelium, skin, and muscle, respectively. From their pri- Note that humans are highly polymorphic for the
mary sites of infection, these viruses can infect peripheral genes that encode MHC molecules. That is, there are nu-
neurons that innervate those sites. Those neurons then merous alternative forms, or alleles, of the individual
provide a conduit to the central nervous system. MHC genes in the human population. The advantages of
that polymorphism are discussed later. For the moment,
Cell-Mediated Immunity note the following. A T-cell receptor recognizes a particu-
Whereas the humoral immune response generates anti- lar antigenic peptide only when it is associated with an
bodies that interact with and neutralize extracellular vi- MHC molecule that is encoded by a particular MHC al-
ruses in the blood and tissues, CTLs of the cellular arm of lele. That is, a T-cell receptor recognizes the complex of a
the adaptive immune response recognize and kill virus- particular antigenic peptide and a particular MHC mole-
infected cells. To understand how CTLs might carry out cule. Consequently, a T cell that recognizes a peptide pre-
these crucial functions, we need to answer at least four sented by a particular MHC allotype will not recognize
basic questions. First, by what means does an infected cell that same peptide when it is presented by a different MHC
alert the effectors of adaptive immunity that it is, in fact, allotype. This phenomenon is known as MHC restriction
infected? Second, by what means are the effector cells of (Boxes 4.9 and 4.10).
the adaptive response able to recognize that a cell is in- A typical T-cell receptor is comprised of an chain
fected? Third, how do the effector cells of adaptive immu- and a chain, both of which are anchored to the plasma
nity destroy an infected target cell? Pertinent to all three membrane by a transmembrane domain (Fig. 4.13). Like
questions is the fact that all the effectors of adaptive im- the heavy and light chains that make up an antibody mol-
munity, humoral as well as cellular, are specific for their ecule, the and chains of the T-cell receptor contain a
particular cognate antigens. So, fourth, by what means are variable region and a constant region. As you might have
the mediators of the cellular response able to specifically surmised, the variable regions of T-cell receptors contain
recognize any of the particular innumerable antigenic de- the binding sites for the complexes of MHC molecules
terminants that might be generated by any of the many and antigenic peptides.
different pathogens that might infect a cell? The varied T-cell receptors are generated by modular
The following points are relevant to the first two of the design, in much the way that the diverse B-cell receptors
above questions. Whereas B-cell receptors and soluble an- and antibodies are generated. That is, the variable regions
tibodies recognize specific epitopes on the whole patho- of T-cell receptor chains are encoded in a series of sets of
gen or on whole proteins of the pathogen, T-cell receptors separate gene segments. Alternative forms of each seg-
recognize particular short peptides derived from the ment are present on the DNA, and they can be recom-
pathogens proteins. Most importantly, T-cell receptors bined in many different combinations. During recombi-
recognize those pathogen-derived peptides only when nation, additional nontemplated nucleotides are inserted
they have been assembled into a complex with MHC mol- to generate greater T-cell receptor diversity. As in the case
ecules, which are then expressed at the surface of the in- of the gene segments making up B-cell receptors and an-
fected cell. That is, MHC molecules are integral mem- tibodies, the segments forming T-cell receptors are re-
brane glycoproteins that present antigenic peptides at the ferred to as V, D, J, and C segments. As then might be ex-
cell surface, where they might then be seen by effector pected, each mature T cell produces only one functional
cells of cell-mediated immunity. chain and one functional chain, which together make
The interactions between a T-cell receptor and an up its T-cell receptor. (Note that rearrangements of heavy-
MHC molecule/peptide complex are shown schemati- and light-chain gene loci, which result in B-cell receptors
cally in Fig. 4.13. Notice how the components of the and antibodies, occur only in B cells and that rearrange-
T-cell receptor make contact with both the antigenic pep- ments at and chain loci, which give rise to T-cell re-
tide and the MHC molecule. Antigen presentation by ceptors, occur only in T cells.) Importantly, T cells, like
114 CHAP TER 4

Figure4.13 The left panel depicts a T-cell recep- CD8 binds the 3 domain of MHC class I CD4 binds the 2 domain of MHC class II
tor on a CTL interacting with an MHC class I mol-
ecule/antigenic peptide complex on a target cell. Target cell Antigen-presenting cell
The right panel depicts a T-cell receptor on a
helper T cell interacting with an MHC class II mol-
ecule/antigenic peptide complex on an antigen-
presenting cell. Notice how the components of the
2-micro- 3 CD8 2 2 CD4
T-cell receptor contact the antigenic peptide, as
well as the MHC molecule. The CD8molecule on globulin
2 1 1
the CTL interacts with the 3 domain of the MHC
class I molecule, ensuring that MHC class I mole-
cules present peptides to CD8 cells only (left panel). T-Cell
Similarly, the CD4molecule on the helper T cell receptor
interacts with the 2 domain on the MHC class II
molecule, ensuring that MHC class II molecules
present antigenic peptides to CD4 cells only (right
panel). Adapted from P. Parham, The Immune
System, 2nd ed. (Garland Science, New York, NY, CD8 T cell CD4 T cell
2005), with permission.

Box 4.9 Box 4.10

MHC restriction was discovered by Rolf Zinkernagel MHC molecules play an essential role in adaptive im-
and Peter Doherty in 1974 at the John Curtin School of munity. What then is the meaning of major histocom-
Medical Research in Canberra, Australia. The discovery patibility complex? The answer is as follows. Because
was made while Zinkernagel and Doherty were trying of the high level of polymorphism at the MHC locus
to explain how LCMV, an arenavirus, although noncy- in humans, there are an exceptionally large number of
topathic, nevertheless gives rise to lethal infections in combinations of antigenically distinct MHC molecules
the brains of mice. They hypothesized that death invivo expressed by different individuals in the human popu-
was due to immunopathology resulting from CTLs lation. The differences between the MHC molecules
attacking infected brain. In one experiment to test this expressed by different individuals are the major cause
hypothesis, they isolated CTLs from infected mice and of graft rejection of transplants from unrelated donors.
examined whether they might attack LCMV-infected Historically, the MHC gene complex was recognized for
cells in culture. Their key observation was that these this graft rejection phenomenon before its actual physi-
CTLs indeed destroyed LCMV-infected cells invitro ological function in immunity was understood.
but, unexpectedly, only if the target cells expressed the
same MHC isoform as the CTLs. The CTLs did not at-
tack uninfected cells, regardless of their MHC isoform,
nor did they attack cells infected with an unrelated B cells, are amplified by clonal selection, with different
virus. Thus, a T-cell receptor recognizes a specific T-cell clones recognizing different antigens.
antigenic peptide in the context of a particular MHC As might be expected from the above, there are notable
isoform. This study provides yet another example of similarities between T-cell receptors and B-cell receptors.
virology-based studies leading to major breakthroughs For example, the and chains of T-cell receptors are
in other disciplines, in this instance in immunology. It each folded into conformations that resemble Ig domains,
was a wonderful example of how certain things cannot and the three-dimensional structures of their antigen-
be planned, said Zinkernagel. Absolutely, this was a binding regions resemble those of antibodies.
miracle of chance. Despite the similarities between T-cell receptors and
Reference B-cell receptors, they differ in important ways. For exam-
Zinkernagel, R. M., and P. C. Doherty. 1974. Restriction of ple, T-cell receptors do not undergo mutation or class
in vitro T-cell-mediated cytotoxicity in lymphocytic chorio switching after gene rearrangement. The latter is so be-
meningitis within a syngeneic or semiallogeneic system. Nature
(London) 248:701702. cause there is only one C segment and two functionally
equivalent C segments.
Host Defenses and Viral Countermeasures 115

Another important difference between T-cell receptors where they can interact with each other when they do so.
and B-cell receptors is that there is no soluble form of the T- Interactions between these lymphocytes are important
cell receptor. T-cell receptors function solely as cell surface because B cells, as well as CTLs, generally must interact
receptors for MHC molecule/antigenic peptide complexes. with Th cells to be activated and undergo clonal expan-
Despite the facts that T-cell receptor genes are assem- sion. Also note that lymph nodes are anatomically orga-
bled by modular design and that T cells undergo clonal nized to facilitate the interactions between Th cells and
selection, T-cell receptors also differ from B-cell receptors B cells and CTLs. Other Th cells leave the lymph nodes
and antibodies with respect to their degree of degeneracy. and travel via the efferent lymph and blood to the infected
That is, T-cell receptors are less specific for a particular tissue, where they act together with macrophages to en-
antigen than are B-cell receptors, and thus are more likely hance the inflammatory response.
than B-cell receptors to cross-react with unrelated anti- Here is another important particular. As there are two
gens. The reason for the degeneracy of the T-cell receptor classes of T cells, likewise, there are two classes of MHC
is that all T-cell receptors recognize the same common molecules. MHC class I molecules present or display en-
structure of a peptide situated in the groove of an MHC dogenous antigenic peptide fragments. That is, they pres-
molecule (Fig. 4.13). The structure of the groove is de- ent antigenic peptides that have been generated from pro-
scribed in detail below. For now it is sufficient to note that teins that are produced by the pathogen during its replication
the MHC/peptide complex appears as a relatively flat sur- cycle in the cell displaying those peptides. In contrast, MHC
face to the T-cell receptor. Moreover, the bound peptide class II molecules present exogenous antigenic peptides. That
represents less than one-third of that surface. In contrast, is, they present peptides that are generated by special phago-
B-cell receptors and antibodies recognize antigens that are cytic cells, which phagocytose the pathogen and process it for
not submerged within the context of an antigen-presenting presentation by MHC class II molecules. These special
MHC molecule. phagocytic cells are the professional antigen-presenting cells,
T-cell cross-reactivity is not necessarily disadvanta- which include dendritic cells, macrophages, and B cells. Their
geous to the host. For example, it can be beneficial in the job as professional antigen-presenting cells is to alert Th cells
case of a T-cell receptor that cross-reacts with serological that an infection is under way.
variants of a particular virus, such as influenza virus The above seemingly disjointed facts come together as
(Chapter 12). Cross-reactive T cells also account, at least follows. Virtually all cells in the body express MHC class I
in part, for the finding that earlier virus infections can molecules, which present endogenous antigenic peptides
have a profound effect on a hosts response to later infec- that inform CTLs when a cell is infected and should be
tions with unrelated viruses. For example, mice that are destroyed. In contrast, only professional antigen-presenting
immune to LCMV clear Pichinde virus or vaccinia virus cells express MHC class II molecules, which present exog-
faster than mice not immune to LCMV. This enhanced enous antigenic peptides to inform Th cells that an infec-
clearance of the unrelated viruses is due to cross-reactive tion is under way. When so informed, Th cells can take up
LCMV-specific memory CTLs. On the other hand, cross- their role as organizers of the adaptive immune response.
reactive T cells can be harmful if they lead to immuno (Neurons do not express MHC class I molecules, as noted
pathology, resulting from T cells cross-reacting with tis- above. Erythrocytes, which do not contain nuclei, also do
sues of the host (see below). not express MHC class I molecules. Perhaps this is the fac-
Now here is a very important proviso to our story. tor that predisposes them to persistent infection with ma-
There are two functionally distinct classes of T cells, each larial parasites.)
of which expresses T-cell receptors. Thus far, we have Since T-cell receptors recognize an antigen in the form
mentioned only the CTLs that directly attack and destroy of a peptide fragment in association with an MHC mole-
virus-infected cells. In addition, there are helper T cells cule, how are these antigenic peptides generated, and how
(Th cells), which orchestrate the adaptive immune re- are they coupled with MHC molecules so that they might
sponse. They do so by secreting the particular cytokines be presented to T cells? These processes are described in
that the particular situation calls for, in that way activat- some detail below. However, before considering them, we
ing either CTLs, or B cells, or both. How Th cells might first need to consider some other issues. We begin by having
know what is needed is described below. But first, there a more detailed look at the structures of MHC molecules.
are several other points to note. MHC class I molecules are comprised of a membrane-
Some Th cells circulate in the blood, while others con- bound heavy or chain and a noncovalently bound
gregate in lymph nodes. The reason why some Th cells are chain, also called 2-microglobulin. The chain contains
found in lymph nodes is because that is where Th cells three extracellular domains: 1, 2, and 3. The two most
and B cells often first encounter their cognate antigen and membrane-distal domains, 1 and 2, fold together to
116 CHAP TER 4

MHC class I molecule MHC class II molecule

Peptide- Peptide-
binding binding
groove groove

2 1 1 1

3 2-micro- 2 2

Figure4.14 MHC class I molecules (left panels) are comprised of a

Peptide- Peptide-
membrane-bound heavy or chain and a noncovalently bound chain,
binding binding
also called 2-microglobulin. The chain contains three extracellular
groove 1 groove
domains: 1, 2, and 3. The two most membrane-distal domains, 2 1 1
1 and 2, fold together to constitute the deep peptide-binding groove.
MHC class II molecules (right panels) are comprised of two membrane-
bound polypeptide chains, referred to as the and chains. Each of the
MHC class II chains contains two extracellular domains: 1 and 2, and 1
and 2, respectively. The most membrane-distal (1 and 1) domains
come together to form the peptide-binding groove. The 3 domain and
2-microglobulin of MHC class I molecules, as well as the 2 and 2 do-
mains of MHC class II molecules, are Ig-like domains. In each instance,
they form the stem structures that support the peptide-binding domains
of the respective MHC class molecules. Thus, MHC class I and MHC class
II molecules each contain two similar domains that come together to form 2-micro- 2
a peptide-binding groove, which sits on two Ig-like domains. Adapted globulin 2
from P. Parham, The Immune System, 2nd ed. (Garland Science, New York,
NY, 2005), with permission.

comprise the deep peptide-binding groove (Fig. 4.14 1) domains come together to form the peptide-binding
and 4.15). groove on MHC class II molecules.
MHC class II molecules are comprised of two mem- The 3 domain and 2-microglobulin of MHC class I
brane-bound polypeptide chains, referred to as the and molecules, as well as the 2 and 2 domains of MHC class
chains (Fig. 4.14 and 4.15). Each of the MHC class II II molecules, are Ig-like domains. In each instance, they
chains contains two extracellular domains: 1 and 2, and form the stem structures that support the peptide-binding
1 and 2, respectively. The most membrane-distal (1 and domains of the respective MHC molecules. Thus, MHC

Figure4.15 Looking down at the top of an MHC class I peptide-binding groove (left panel) and an
MHC class II peptide-binding groove (right panel), notice how the structure of the MHC class II
peptide-binding groove resembles that of MHC class I molecules. In each instance, the sides of the
groove are comprised of helices and the floors are formed by a sheet of eight antiparallel strands.
A bound peptide is shown in each instance. Peptides are bound to MHC class I molecules by interac-
tions at their ends and to MHC class II molecules by interactions along their length. Adapted from
P. Parham, The Immune System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.


Peptide-binding groove


helix groove
2 1
MHC class I MHC class II
Host Defenses and Viral Countermeasures 117

class I and MHC class II molecules each contain two sim- Although the peptide-binding grooves of MHC class I
ilar domains that come together to form a peptide-binding and MHC class II molecules resemble each other in form,
groove, which sits on two Ig-like domains. Note how the there are differences between them. One of the most im-
structure of the MHC II peptide-binding groove resem- portant differences lies at the ends of the grooves. These
bles that of MHC I molecules. In each instance, the sides ends are more open in the case of MHC class II molecules.
of the groove are comprised of helices, and the floors are As a result, peptides that bind to MHC class I molecules
formed by a sheet of eight antiparallel strands. are usually 8 to 10 amino acids long (restricted in length
In contrast to the genes that encode T-cell and B-cell re- by the ends of the groove), whereas peptides that bind to
ceptors and antibodies, the genes of the MHC complex are MHC class II molecules are at least 13 amino acids long and
conventional and do not undergo rearrangement. Also, in may be much longer. In each instance, the peptide lies in an
contrast to the specificity displayed by T-cell and B-cell re- elongated conformation within the groove (Fig. 4.15).
ceptors and antibodies, each MHC molecule can present a Now, consider the following puzzler. On the one hand,
variety of antigenic peptide fragments. Yet no MHC mole- a particular MHC isoform must be able to present a vari-
cule can present all antigenic peptide fragments. How then ety of antigenic peptide fragments. Moreover, it must be
might an individual be able to display antigenic peptide able to stably bind each of those antigenic peptides. If that
fragments from whatever pathogen might be encoun- were not the case, then peptide exchanges could occur at
tered? The answer is that each individual has six different the cell surface, in that way compromising the ability of
genes for the MHC class I chain, which forms the pep- MHC molecules to accurately reflect the events taking place
tide-binding groove on MHC class I molecules. Impor- within a cell. It would be especially troubling if MHC class
tantly, with respect to the population per se, three of these I molecules at the surface of an uninfected cell were to ac-
genes, HLA-A, HLA-B, and HLA-C, are highly polymor- quire antigenic peptides released from an infected cell. If
phic (Box4.11). That is, there are 218, 439, and 96 known that were to happen, then that uninfected cell might be at-
variants of HLA-A, HLA-B, and HLA-C, respectively, in the tacked by CTLs. That is, the uninfected cell might become
human population. Moreover, an individual is likely to be an innocent bystander. On the other hand, this ability of
heterozygous at each of these loci. (The 2-microglobulin is individual MHC isoforms to stably bind a variety of pep-
monomorphic. Regardless, it does not contribute to the tides is very different from the behavior of peptide-binding
peptide-binding site.) Likewise, there is a high level of poly- receptors (e.g., hormone receptors), which usually are spe-
morphism for several of the genes that encode the MHC cific for just one peptide. So, how then do MHC isoforms
class II and chains, which together form the peptide- manage to stably bind a variety of different peptides?
binding groove on MHC class II molecules. The rather elegant solution conceived by Mother Nature
Each variant or isoform of an MHC molecule is differ- is as follows. The peptides that bind to any particular
ent with respect to the antigenic peptides it can present to MHC isoform have the same or very similar amino acids
T-cell receptors. Consequently, the relatively large num- at two or three specific positions within the chain. There
ber of MHC alleles within each individual and the high are pockets within the groove of that MHC molecule that
level of polymorphism within the population ensure that can accommodate those specific amino acids on the pep-
many individuals, and the population in general, can re- tide. Importantly, the amino acids that line those particular
spond to virtually any pathogen that might be encountered. pockets in the groove are the very ones that are highly poly-
(What might be the implications of these points with re- morphic within the population (Fig. 4.16). Thus, the groove
gard to highly inbred human and animal populations?) is specific for only a few sites on the antigenic peptide, and
several variants of the groove are present in every individ-
Box 4.11 ual. Moreover, as is discussed below, the antigenic peptide
actually stabilizes the dimeric MHC molecule. Were it to be
lost at the cell surface, the MHC molecule would disassem-
The human MHC is also referred to as the human
ble and consequently not be able to acquire an exogenous
leukocyte antigen (HLA) complex because the proteins
it encodes are found on leukocytes (white blood cells).
peptide. Furthermore, disassembled MHC class I molecules
MHC (HLA) proteins should not be confused with the are in fact released from the cell surface by proteolytic cleav-
perhaps better known ABO cell surface antigens, which age, further ensuring against innocent bystander events.
are found on red blood cells. An individuals ABO type is Here is an important point that has not yet been explic-
determined by the inheritance of 1 of 3 alleles from each itly stated. The T-cell receptors of CTLs and Th cells are
parent (A, B, or O), and these do not change as a result of exactly the same. That is, they are generated from the very
environmental influences. same genomic sequences. Indeed, T-cell receptors are gen-
erated before T-cell progenitors decide to become CTLs
118 CHAP TER 4

MHC class I variability or Th cells. Next, let us recall that MHC class I molecules
present endogenous foreign peptides for recognition by
2 1 CTLs, whereas MHC class II molecules present exogenous
foreign peptides for recognition by Th cells. How then are
3 T-cell receptors of CTLs able to be specific for MHC class
I molecules, and how are T-cell receptors of Th cells able
to be specific for MHC class II molecules? The answer is
that T cells express coreceptor molecules, referred to as
CD4 and CD8. Th cells generally express CD4 corecep-
tors, and CTLs generally express CD8 coreceptors. Indeed,
MHC class II variability Th cells are often called CD4 T cells, and CTLs are often
called CD8 T cells. The CD4 and CD8 coreceptors are
responsible for the specificity of Th cells for MHC class II
1 1
molecules and of CTLs for MHC class I molecules, re-
2 2 spectively. The coreceptors confer this specificity by cou-
pling to their respective cognate MHC molecule, thereby
strengthening the adhesion between the T-cell receptor
and the appropriate class of MHC molecule (Figure 4.13).
Consequently, CTLs can focus their attention on virus-
infected cells, and Th cells can get their cues from profes-
Figure4.16 MHC molecules bind peptides at specific sites
sional antigen-presenting cells (Boxes 4.12 and 4.13).
within the peptide-binding groove, as highlighted by the red dots When T-cell receptors of a CTL bind to MHC class I
(top, MHC class I; bottom, MHC class II). The amino acids that molecules on an infected cell, a signal is transmitted into
constitute those particular sites are the very ones that are highly poly- the CTL informing it to kill. When T-cell receptors of a
morphic within the population. Thus, the groove is specific for only
a few sites on the antigenic peptide, and several variants of the groove
Th cell bind to MHC class II molecules on a professional
are present in every individual. Adapted from P. Parham, The Immune antigen-presenting cell, a signal is transmitted into the
System, 2nd ed. (Garland Science, New York, NY, 2005), with Th cell, activating it to provide help (i.e., to help activate a
permission. B cell or CTL). Yet as we noted above, the T-cell receptors

Box 4.12
Biological systems often turn out to be considerably more In addition to evidence suggesting that CD4 CTLs might
complex than our sometimes simplistic working models act to promote the clearance of virus infections, there also
would have them be. In this particular instance, some CD4 is evidence that these cells might account in part for the
T cells are known to express CTL activity, in addition to immunopathology associated with many, if not most, virus
their better-known helper activity. Indeed, CD4 and CD8 infections (see Immunopathology in this chapter). LCMV
CTLs against chicken pox and parainfluenza viruses exist in provides a well-known example of this phenomenon. In
equal frequencies, and up to 50% of the CD4 T cells in some LCMV-infected mice that are depleted of CD8 T cells, pas-
HIV-infected individuals display clear cytotoxic potential. sive transfer of MHC class II-restricted CD4 CTLs results in
LCMV disease.
Although the contribution of CD4 CTLs to antivirus de-
fenses invivo is not yet clear, there is some indication that Also note that whereas CD4 T cells are better known as
these cells might play a role in the clearance of HSV from cytokine producers than are CD8 T cells, much CD8 T-cell
recurring herpetic lesions. For example, CD4 T cells are de- activity is mediated by IFN-, in addition to CTL-specific
tected in herpetic lesions at a time when viral titers are de- killing mechanisms (see the text). Perhaps we should be
clining, whereas CD8 T cells are found in these lesions only thankful that CD8 T cells are not known to express helper
several days later, after virus titers have already declined. activity, since that would complicate matters even further.
Host Defenses and Viral Countermeasures 119

Box 4.13 The ubiquitin-proteasome pathway begins with pro-

teins (both self proteins and ones encoded by an intracel-
lular pathogen) being marked for degradation by ubiq-
As you may already know, CD4molecules are the major
uitin, a 76-amino-acid polypeptide that the cell attaches
cell surface receptor for HIV. That fact is the reason why
HIV specifically targets CD4 Th cells. Importantly, the
to the amino groups on lysine residues of proteins des-
tropism of HIV for CD4 T cells dramatically shapes the tined for destruction (Fig. 4.17). Ubiquitination is the
clinical course of AIDS. That is so because CD4 Th cells principal means by which a cell marks any protein for
have a crucial role in orchestrating the activities of essen- degradation. Additional ubiquitins are added to form a
tial immune effector cells, including CD8 CTLs, B cells, long polyubiquitin chain. The polyubiquitinated protein
and macrophages, doing so by directly interacting with then is recognized by a proteasome, a large barrel-shaped
them and by the cytokines they secrete. The loss of CD4 protein complex comprised of some 28 subunits, which
T cells in AIDS results in the severest form of immuno- expresses several different protease activities and degrades
deficiency known. the protein into polypeptides.
The peptides that are generated by the proteasome are
released into the cytosol. Those peptides that are 9 amino
acids long or longer are then transported into the lumen
are the same on each of these cell types. How then are these of the endoplasmic reticulum (ER) by the protein com-
identical T-cell receptors able to transmit distinctly different plex referred to as the transporter associated with antigen
signals? The answer is that CD4 and CD8 are also signaling processing (TAP), which is embedded in the ER mem-
molecules and the signals they transmit are likely quite dif- brane. TAP is actually a heterodimer, comprised of TAP-1
ferent. Thus, the respective signals transmitted by CD4 and and the related TAP-2 (Box4.14).
CD8 are probably in part responsible for the functional Since MHC class I chains are transmembrane pro-
differences between CTLs and Th cells. Other functional teins, they contain signal sequences that target them to the
differences between CTLs and Th cells may also affect the ER membrane while they are still ribosome-bound na-
signals transmitted by their respective T-cell receptors. scent polypeptides (see Chapter 2 for the generation of
transmembrane proteins). Likewise, 2-microglobulin is
Antigen Presentation by MHC Class I targeted to the ER. Within the ER, the newly synthesized
Molecules MHC class I chains and 2-microglobulins associate with
Recall that virtually all cells express MHC class I molecules each other and with peptide fragments generated in the cy-
and that these molecules serve to present endogenously tosol by the proteasome. Then, the complex of MHC class I
synthesized peptides to CTLs. The process of antigen pre- molecule and peptide is transported by the usual vesicle-
sentation by MHC class I molecules thus begins with the mediated secretory pathway to the plasma membrane.
degradation of endogenously synthesized viral proteins. Before moving on, there are several points that bear
In this regard, eukaryotic cells have two major protein reiterating. First, the MHC class I peptide presentation
degradation pathways. The ubiquitin-proteasome path- pathway occurs in noninfected as well as infected cells.
way mediates degradation of endogenously synthesized Moreover, it presents host- as well as pathogen-derived
proteins, whereas exogenous proteins are degraded in lys- peptides. Because or in spite of this fact, it informs CTLs
osomes. Thus, MHC class I molecules present peptides as to whether or not the cell might be infected. Second,
generated by proteasomal degradation of endogenously nearly all cell types are able to express MHC class I mole-
synthesized viral proteins. cules. Here then is a new point. IFN-/ greatly increase
In actuality, a fraction of all the proteins made in the cell transcription of MHC class I genes and of the genes encod-
are degraded by the ubiquitin-proteasome pathway, regard- ing the TAP transporter proteins and the LMP2, LMP7, and
less of whether the cell is infected. Moreover, some of the MECL-1 proteasome subunits (Box4.14). Thus, IFN-/
peptides that are thus generated, host as well as viral, are enhance the efficiency of the MHC class I antigen presenta-
presented by MHC class I molecules at the cell surface. Im- tion pathway, in that way causing a virus-infected cell to be
portantly, if the cell is infected, then some of the peptides more discernible to CTLs, another example of the interplay
presented at the cell surface will be derived from virus- between the innate and adaptive immune systems.
encoded proteins that were in the cytosol before the emer-
gence of progeny virus particles. That is a key point, since it Antigen Presentation by MHC Class II
means that a cell can be recognized and killed by CTLs, in Molecules
the early stages of infection of that cell, before any progeny We begin our consideration of antigen presentation by
infectious virus particles might be assembled within it. MHC class II molecules by recalling that virtually all cells
120 CHAP TER 4

2m Lumen
of ER

MHC class I Tap1 Tap2


Medial Golgi apparatus




Transport vesicles

Presenting cell

T cell receptor

T cell

Figure4.17 The pathway for the processing and presentation of endogenous proteins by MHC class I proteins begins
with both self proteins and ones encoded by an intracellular pathogen being marked for degradation by ubiquitin, a
76-amino-acid polypeptide that the cell attaches to the amino groups on lysine residues of proteins destined for de-
struction. Additional ubiquitins are added to form a long polyubiquitin chain (not shown). The polyubiquitinated
protein then is recognized by a proteasome, a large barrel-shaped protein complex comprised of some 28subunits,
which expresses several different protease activities and degrades the protein into polypeptides. The peptides that are
generated by the proteasome are released into the cytosol. Those peptides that are nine or more amino acids long are
transported into the lumen of the ER by the protein complex referred to as the transporter associated with antigen
processing (TAP), which is embedded in the ER membrane. Since MHC class I chains are transmembrane proteins,
they contain signal sequences that target them to the ER membrane while they are still ribosome-bound nascent poly-
peptides. Likewise, 2-microglobulin is targeted to the ER. Within the ER, the newly synthesized MHC class I chains
are stabilized by the membrane chaperone protein calnexin, until the MHC class I chains associate with 2-microglobulin.
Next, the peptide fragments generated in the cytosol by the proteasome bind to the MHC class I molecule, and the
MHC class I/peptide complex is transported by the usual vesicle-mediated secretory pathway to the plasma mem-
brane. At the cell surface, the MHC class I/peptide complexes may then interact with the T-cell receptors of a cytotoxic
CD8 cell. Adapted from S. J. Flint, L. W. Enquist, R. M. Krug, V. R. Racaniello, and A. M. Skalka, Principles of
Virology: Molecular Biology, Pathogenesis, and Control, 2nd ed. (ASM Press, Washington, DC, 2003), with permission.
Host Defenses and Viral Countermeasures 121

Box 4.14 synthesized at the ER and are transported to the plasma

membrane via the secretory pathway (Fig. 4.18). Thus,
evolution had to devise a means to prevent MHC class II
Here is yet another example of the elegance of the mam-
molecules from acquiring endogenous peptides that are
malian immune response. First, bear in mind that a
major purpose of the ubiquitin-proteasome pathway is
generated in the cytosol by the ubiquitin-proteasome
to degrade faulty or misfolded cellular proteins. Sec- pathway and transported into the ER by the TAP trans-
ond, when an infection is under way, IFN- induces the porter. To avoid this inappropriate loading of endogenous
expression of three proteins, LMP2, LMP7, and MECL-1, peptides onto MHC class II proteins, professional anti-
which replace three constitutive components of the gen-presenting cells express a molecule referred to as the
proteasome. The incorporation of these IFN--inducible MHC class II-associated invariant chain (Ii), which binds
components into the proteasome modifies its specific- to partially folded MHC class II molecules in the ER. Ii
ity, such that it generates peptides that contain residues binds to MHC class II molecules in such a way that part of
that are preferred by the anchoring sites of the peptide- Ii is in the peptide-binding groove, thereby blocking any
binding grooves of MHC class I molecules. Moreover, the interaction of MHC class II molecules with endogenous
incorporation of these IFN--inducible components into peptides in the ER.
the proteasome also causes it to generate peptides that Evolution is not only clever but also efficient, for Ii also
are the preferred size for transport into the ER by TAP facilitates the export of MHC class II molecules from the
(see the text). ER. Moreover, Ii also targets MHC class II molecules to
the endosomal/lysosomal compartment, where exogenous
peptides are generated and loaded onto MHC class II
molecules. This Ii-mediated targeting of MHC class II
express MHC class I molecules and that MHC class I mol- molecules is noteworthy because the secretory pathway
ecules present endogenously synthesized foreign peptides does not usually intersect the endocytic pathway. Before
to CTLs, so that infected cells might be recognized and loading of exogenous peptides can happen, MHC class II-
killed by CTLs. In contrast, MHC class II molecules are associated Ii is first cleaved and released by the action of
expressed only by professional antigen-presenting cells, lysosomal proteases.
that is, macrophages, dendritic cells, and B cells. More- One additional point about antigen presentation by
over, the function of MHC class II molecules is to present MHC class II molecules is that MHC class II molecules
peptides generated from exogenous (i.e., endocytosed) are generated and intersect the endosomal/lysosomal
proteins, in order to alert Th cells that an infection is under compartment in noninfected as well as infected cells. Con-
way in the host. As we will see, these Th cells then activate sequently, like MHC class I molecules, MHC class II mol-
other immune effectors. For example, they stimulate ecules also present self peptides. This brings up what is
B cells to produce antibodies and may help activate CTLs. perhaps the most intriguing aspect of the mammalian im-
Importantly, Th cells also secrete cytokines (e.g., IFN-), mune response: how does it distinguish nonself from self
which enhance the expression of MHC molecules and other (see below)? However, there are several other issues that
components of the antigen presentation machinery. need to be considered first.
As noted above, the ubiquitin-proteasome pathway,
which mediates the degradation of endogenously synthe- The Rationale for MHC Restriction
sized proteins for MHC class I peptide presentation, is Presentation of antigen by MHC molecules is certainly el-
one of the two major protein degradation pathways in eu- egant, but is it a needlessly complex process? That is, why
karyotic cells. The other is lysosomal proteolysis, by which not have CTLs simply recognize antigen that is presented
exogenous proteins that have been taken up by endocyto- in a non-MHC-restricted manner, the way B-cell recep-
sis are degraded. Lysosomes contain several proteases for tors do? We cannot know precisely what evolution had in
this purpose. With these points in mind, professional mind when concocting MHC restriction, but a reasonable
antigen-presenting cells acquire exogenous antigenic pro- supposition is that MHC restriction enables CTLs to
teins by endocytosing whole pathogens and generate anti- focus their attention on virus-infected cells, without
genic peptides from them by lysosomal proteolysis. having their potent and energetically costly effector func-
Before we consider the pathway for MHC class II pep- tions sidetracked by free extracellular virus particles or
tide presentation, we first need to take into account the proteins. Antibodies can deal well enough with free vi-
following. Since MHC class II molecules, like MHC class I rus, and they are much more plentiful and cost-effective
molecules, are transmembrane glycoproteins that are for that purpose. Recall that each plasma B cell can pump
expressed at the plasma membrane, they likewise are out thousands of antibody molecules every second! Yet
122 CHAP TER 4


Antigen- TH cell
presenting cell

pH 7.5
Early endosome

Late endosome
pH 5.5
class II

T cell
Fusion receptor
Invariant Invariant chain vesicle MHC class II
chain degrades vesicle
pH 4.5
Golgi apparatus


Figure4.18 The MHC class II peptide presentation pathway presents peptides generated by lysosomal proteolysis of
endocytosed pathogens and proteins. MHC class II molecules are synthesized and assembled in the ER. During their
transport to the plasma membrane via the secretory pathway, they are diverted to the internalized endosomes that
contain the pathogen-derived peptides. The secretory vesicles containing the MHC class II molecules then fuse with
these endosomes, and the MHC class II molecules acquire the peptides that they present at the cell surface. MHC class II
molecules are prevented from acquiring endogenous peptides, which are generated in the cytosol by the ubiquitin-
proteasome pathway and transported into the ER by the TAP transporter, as follows. Professional antigen-presenting
cells express a molecule referred to as the MHC class II-associated invariant chain (Ii), which binds to partially folded
MHC class II molecules in the ER in such a way that part of Ii is in the peptide-binding groove, thereby blocking any
interaction of MHC class II molecules with endogenous peptides in the ER. Ii also facilitates the export of MHC class II
molecules from the ER. Moreover, Ii also targets MHC class II molecules to the endosomal/lysosomal compartment,
where loading of exogenous peptides occurs. This Ii-mediated targeting is noteworthy because the secretory pathway
does not usually intersect the endocytic pathway. Before loading of exogenous peptides can happen, MHC class II-
associated Ii is first cleaved and released by the action of lysosomal proteases. At the cell surface, the MHC class II/
peptide complex is available to the T-cell receptor of CD4 Th cells. Adapted from S. J. Flint, L. W. Enquist, R. M. Krug,
V. R. Racaniello, and A. M. Skalka, Principles of Virology: Molecular Biology, Pathogenesis, and Control, 2nd ed. (ASM
Press, Washington, DC, 2003), with permission.

those antibodies cannot reach intracellular virus, whereas antigens bound to uninfected cells, the result would be the
the vastly more elaborate CTLs can. Why then have those innocent-bystander effect mentioned earlier.
very special CTLs distracted by extracellular virus? The above two arguments pertain to antigen presenta-
Here is another important reason for MHC restriction. tion by MHC class I molecules, which results in CTLs rec-
Suppose that CTLs did recognize non-MHC-associated ognizing and attacking infected cells. But why have MHC
antigens at the cell surface. Then, consider the consequences restriction in the case of MHC class II molecules, which
if an uninfected cell were to have viral antigens bound present antigen to Th cells? A plausible answer is based
to it. Such antigens might be contained in the debris on three points. First, Th cells must be activated to con-
from neighboring infected cells. If CTLs were to recognize tribute to the immune response, and second, professional
Host Defenses and Viral Countermeasures 123

antigen-presenting cells must deliver two signals to Th presentation, as well as a costimulatory signal via B7, is a
cells in order to activate them. Before considering the defining property of these cells. Indeed, that is why they are
third point, note that the above two points play out as fol- uniquely endowed to activate Th cells.)
lows. The first signal from the professional antigen- Now here is the third point. In noninfected tissues,
presenting cell is triggered by protein B7, on its surface, professional antigen-presenting cells are in a resting
which interacts with protein CD28 on the surface of the state, in which they express very low levels of MHC class II
Th cell (Fig. 4.19). Note that the signal triggered by B7 is molecules and of B7. However, when an infection is un-
nonspecific, in the sense that CD28 on any Th cell might der way, inflammatory cytokines such as IFN- are
respond to it. However, the second signal is specific, since produced, which induce the expression of elevated levels
it is triggered by the MHC class II/peptide complex, which of MHC class II molecules and B7. The professional
is recognized specifically by T-cell receptors on particular antigen-presenting cell thus activated can then deliver
T cells. Actually, these two signals may be delivered at the its two signals: a general one in response to the infection
same time. (Note that the capability of professional antigen- and a specific one in response to the particular pathogen.
presenting cells to transmit a stimulatory signal via MHC Herein may lie the rationale for the antigen presentation
by MHC class II molecules expressed by professional anti-
gen-presenting cells. Since the effectors of adaptive immu-
Figure4.19 Professional antigen-presenting cells must deliver two
signals to Th cells in order to activate them. One signal (diagram 1) is nity are powerful, energy-costly, and highly specialized
triggered by the MHC class II/peptide complex on the surface of the weapons, they must not be activated at inappropriate times.
professional antigen-presenting cell, which interacts specifically with That is, they must be activated only in the event of an infec-
T-cell receptors on particular T cells. The second signal (diagram 2) is tion. By making activation of adaptive immunity depen-
triggered by protein B7 on the surface of the professional antigen-
presenting cell, which interacts with protein CD28 on the surface of dent on a nonspecific signal in response to a general infected
the Th cell. The capability of professional antigen-presenting cells to state as well as on a specific signal in response to a particular
transmit a signal via MHC presentation, as well as a costimulatory sig- pathogen, inappropriate activation is not likely to happen.
nal via B7, is a defining property of these cells. The T-cell receptor per Note that a Th cell that receives the two signals then un-
se comprises two proteins, an chain and a chain. The cytoplasmic
tails of the and chains consist of only about three amino acids. dergoes clonal expansion. (Note the analogy between the
Thus, for the T-cell receptor to signal, it is associated with the CD3 dual-key activation system to provide for the failsafe launch-
complex, which comprises four different proteins: one chain, one ing of ballistic missiles and the requirement for two signals
chain, two chains, and two chains. The CD4molecule on the helper to activate affectors of immunity [Box 4.3].)
T cell is shown interacting with the 2 domain on the MHC class II
molecule. Adapted from P. Parham, The Immune System, 2nd ed. Another advantage of peptide presentation via MHC
(Garland Science, New York, NY, 2005), with permission. molecules, apropos both the MHC class I and MHC class II
The co-stimulatory molecule B7 on the
peptide presentation pathways, stems from the fact that
dendritic cell binds CD28 on the naive T-cell peptide fragments rather than whole proteins are displayed.
Since many potential antigenic determinants might be bur-
1 Nave T cell ied within the folded structure of an intact protein, display-
2 ing peptide fragments significantly increases the number of
different epitopes that might be recognized by T-cell recep-
tors. This creates a stronger, more diverse adaptive immune
response. Furthermore, most intact virus proteins are never
present at the cell surface during the viruss replication
cycle. Yet MHC presentation can display epitopes from
essentially every virus protein that is present in the cell.
Relevant to the preceding argument, recall that differ-
ent MHC allotypes contribute different peptide-binding
specificities. Therefore, since each individual is polymor-
B7 phic for MHC class I and MHC class II genes, the number
class II of foreign peptides that can be presented during any
CD4 infection is appreciably amplified. This strengthens the
immune response by increasing the number of activated
pathogen-specific T cells. This argument applies to each in-
dividual within a population. It also applies in an impor-
tant way to the population as a whole. Variability of MHC
Dendritic cell
alleles within a population means that individuals vary in
124 CHAP TER 4

their abilities to respond to particular pathogens, thereby dendritic cells leave the infected tissue and migrate
helping to ensure that at least some individuals survive through the lymphatic system to the nearest draining
any particular epidemic. Small or highly inbred popula- lymph node. This is a key point because nave CTLs and
tions suffer in this regard. Th cells continually circulate from lymph node to lymph
Finally, why have antigenic peptides presented by both node, searching for an antigen-presenting cell that might
MHC class I and MHC class II molecules? Might only one be displaying their cognate antigen. In fact, T cells do not
form of MHC molecule be sufficient? Perhaps, but having leave the circulation unless they are activated by an antigen-
only one form of MHC presentation almost certainly presenting cell which itself has been activated. (Activated
would not be as effective as the system we have. First, if macrophages may provide activating signals to T cells at
there were only one form of MHC molecule recognized the site of infection, to keep them active there until the
by the T-cell receptors of both CTLs and Th cells, then a infection might be cleared.)
professional antigen-presenting cell in the process of acti- Why might the initial activation of Th cells occur in the
vating a Th cell might be the object of attack by a CTL. lymph nodes, rather than at the site of infection? First,
Second, keeping the activation and targeting of Th cells consider that for any given infection, the fraction of cir-
distinct from those of CTLs provides a measure of control culating nave T cells that might be specific for the invad-
that almost certainly benefits the host. ing pathogen might be as few as 1in 106. Thus, there needs
to be some mechanism for rather quickly and efficiently
Activation of Th Cells: Dendritic Cells bringing the minute fraction of nave Th cells which are
and B Cells specific for a particular pathogen into contact with pro-
Before detailing the process of Th cell activation, it would fessional antigen-presenting cells that display that patho-
be good to recap the following points. gens antigens and that can then activate the Th cell. The
As noted before, Th cells must be activated before they migration of dendritic cells from the site of infection to
can carry out their crucial functions. Actually, Th cell activa- the nearest draining lymph node, together with the cir-
tion may be considered to be the first step of a primary adap- culation of nave Th cells from lymph node to lymph
tive immune response. We also emphasized that professional node, is the means by which the immune system solves
antigen-presenting cells must deliver both a stimulatory sig- this problem. This is a rather important concept. Remem-
nal and a costimulatory signal to activate resting or nave Th ber that the adaptive immune system does not attempt to
cells. As also noted, the defining property of the professional initiate an adaptive immune response at the site of infec-
antigen-presenting cells (i.e., macrophages, dendritic cells, tion. Instead, it uses the sentinel dendritic cells to capture
and B cells) is their capacity to transmit a stimulatory signal antigen from the infected site and transport that antigen
via the interaction between an MHC class II/peptide com- to the nearest lymphoid organs, where the effectors of
plex and a T-cell receptor, as well as a costimulatory signal adaptive immunity that might recognize that antigen are
via the interaction between B7 and CD28 (Fig. 4.19). concentrated. Moreover, the fact that T cells do not leave
Also recall that both macrophages and dendritic cells of the circulation unless activated is important with respect
the innate immune system are positioned throughout the to the maintenance of peripheral tolerance (see below).
body, on guard against invading pathogens. However, at un- Once the nave Th cell encounters its cognate antigen-
infected sites, these cells are in a resting state, in which they presenting cell, it matures and undergoes clonal expan-
express low levels of MHC and costimulatory molecules. But sion, a process that takes several days. The activated Th
if the site should become infected, then cytokines are pro- cell then must play its roles helping to activate CTLs and
duced by the innate immune response and activate local B cells. During this entire period, in which first dendritic
macrophages and dendritic cells, causing them to express el- cells are activated at the infected site, followed by activa-
evated levels of MHC molecules and B7. In addition, these tion of Th cells in the lymphoid organs and finally activa-
cells begin to vigorously internalize antigen from the extra- tion of B cells and CTLs, the innate system is doing its best
cellular fluid or from infected cells, which they then can pres- to contain the infection.
ent via MHC molecules. The inflammatory cytokines pro- Here is another detail to note. Th cells tailor an adap-
duced by the innate immune response, together with the tive immune response to the particular situation by se-
encountered antigen, stimulate macrophages and dendritic creting particular combinations of cytokines. For this
cells to mature into fully functional antigen presenters. purpose, nave Th cells can mature into either of two
Dendritic cells are more important than macrophages classes of Th cells, referred to as Th1 and Th2. Th1 cells se-
for initially presenting antigenic peptides to T cells. crete cytokines IFN-, IL-2, and TNF-, which favor a CTL
That is so because macrophages remain at the local site response, whereas Th2 cells secrete cytokines IL-4, IL-5,
of infection, even after being activated. In contrast, activated and IL-10, which favor antibody production.
Host Defenses and Viral Countermeasures 125

There is much research and discussion regarding which recognizes the double-stranded RNA produced during
of these responses is more important in resolving virus infec- infections by a variety of viruses. Thus, dendritic cells are
tions. The answer may vary from one virus to another. Yet the able to recognize a diversity of signals, each characteristic
optimal outcome will likely occur when both arms of adap- of a particular class of pathogen. A second key point is
tive immunity are engaged. Apropos that point, although that dendritic cells can pass this information on to nave
the Th1/Th2 paradigm is useful for the purpose of teach- Th cells when activating them. The dendritic cells do so as
ing, it is a simplification of the actual state of affairs invivo, follows. Whereas the costimulatory signal transmitted by
since the diversity of T-cell responses is more varied than B7 is necessary to activate a nave Th cell, B7 is not the
merely the strict Th1 versus Th2 distinction. That is, T cells only costimulatory molecule expressed by the dendritic
commonly secrete particular combinations of Th1 and Th2 cell. In particular, dendritic cells also secrete cytokines
cytokines. And, in point of fact, there also are other types of which help to activate the Th cell, and the cytokine profile
T cells in addition to Th cells and CTLs (Box 4.15). secreted by the dendritic cell reflects the inputs that the
Whether a nave T cell matures into primarily a Th1 or dendritic cell received at the infected site. The combina-
Th2 cell is partially determined by the costimulation it re- tion of signals transmitted by the dendritic cell when it is
ceives during activation. This occurs as follows. First, at activating the Th cell determines the set of cytokines that
the infected site, dendritic cells receive inputs through the Th cell will then generate. Consequently, the activated
their various cell surface receptors, particularly their Toll- Th cell can, in turn, orchestrate the most effective adap-
like and cytokine receptors, which inform them as to the tive immune response against the invader.
nature of the invading pathogen. For example, recall how B cells also are important antigen presenters during vi-
TLR4senses bacterial lipopolysaccharides, whereas TLR3 rus infections. They carry out this function by specifically

Box 4.15
Regulatory T cells are also part of the T-cell repertoire, serv- Treg cells may be identified by their expression of a tran-
ing to shut down T-cell-mediated immunity towards the scription factor known as FoxP3, which is not known to be
end of an immune response. Regulatory T cells also help to expressed by any other type of cell. Mutations of FOXP3 can
prevent autoimmune reactions, that is, reactions against self. prevent development of Treg cells, resulting in a fatal autoim-
They do this by suppressing autoreactive T cells that may mune disease known as IPEX. Since FoxP3 is encoded by a
have escaped the negative selection process in the thymus gene on the X chromosome, IPEX principally affects boys.
(see the text). The only treatment for IPEX is a bone marrow transplant
from a healthy, HLA-identical sibling.
You may be thinking (mockingly perhaps) that it would
be too simple if there were only one type of regulatory The two known types of adaptive regulatory T cells, Tr1 and
T cell. If so, you indeed are correct. In actuality, there are Th3, are distinguished by their cytokine profiles. Tr1 cells
two known classes of regulatory T cells: the naturally oc- produce high levels of IL-10 and transforming growth factor
curring regulatory T cells (Treg cells) and the adaptive regu- beta (TGF-), whereas Th3 cells primarily produce TGF-.
latory T cells (also known as Tr1 cells and Th3 cells). Each These cells mediate immunosuppression via the secretion
of these regulatory T-cell subsets provides another example of these cytokines. Both IL-10 and TGF- inhibit helper
of the phenomenon whereby one class of lymphocytes can T cells, while TGF- may suppress immunity in general.
regulate the actions of another. The possible relationship between Tr1 and Th3 cells is not
yet clear.
Treg cells appear to originate from a unique lineage of CD4
T cells in the thymus, and they are dedicated to expressing Memory T cells are discussed in the text. Antigen-specific
only a regulatory role. In order to carry out their regulatory memory T cells, which persist long after an infection has
role, Treg cells must make direct contact with responder T resolved, exist for both CD4 Th cells and CD8 CTLs. If at
cells. In contrast, the adaptive regulatory T cells usually start some later time these cells should reencounter their cognate
out as effector T cells that express immunity-promoting antigen, then they are able to quickly activate and multiply.
cytokines. Afterwards, they switch to producing inhibitory They are key components of immunological memory. Three
cytokines. They convey their regulatory effect to responder distinct subpopulations of both CD4 and CD8memory
T cells via these cytokines. T cells are presently known but are not discussed here.
126 CHAP TER 4

binding antigens to their B cell receptors (i.e., cell-bound (Fig. 4.19). One of these signals is the pathogen-specific
antibodies), internalizing them by receptor-mediated en- one, which is triggered by the interaction of the T-cell re-
docytosis and then presenting those antigens via the regu- ceptor on the nave T cell with an MHC class II/peptide
lar MHC class II peptide presentation pathway. However, complex on the professional antigen-presenting cell. The
B cells cannot initiate a primary adaptive response. The other signal is a general one, triggered by the interaction
reason is that nave B cells express only low levels of MHC of CD28 with B7. Since professional antigen-presenting
class II molecules and B7. However, once a B cell has been cells express B7 only in response to inflammatory cyto
activated by an activated Th cell (see below), it expresses kines that are produced when an infection is under way,
greatly increased levels of these molecules. Consequently, B the requirement for the two signals ensures against the in-
cells may not be important antigen-presenting cells during appropriate activation of helper T cells. Consequently, the
the early stages of infection, because their activation is de- B-cell requirement for T-cell help serves as a means to en-
pendent on the prior activation of nave Th cells, which in sure that B cells likewise will be activated only in response
turn depends on the prior activation of dendritic cells. Yet B to an infection.
cells play important roles as antigen-presenting cells during Yet the B cell itself also requires two signals in order to
the later stages of a primary infection, and memory B cells become a plasma cell. In addition to the signal from the
are important antigen presenters during subsequent en- Th cell, the B cell also must recognize its cognate antigen,
counters with the same pathogen. (T-cell-independent acti- further ensuring against its inappropriate activation. In
vation of B cells is possible in some instances [see below].) this regard, recall that virus particles and bacteria are char-
A reason that B cells are especially efficient antigen pre- acterized by regularly repeating surface features (more
senters is that B-cell receptors specifically bind antigen, importantly, epitopes). Indeed, these epitopes are spaced
which is then taken up by endocytosis for processing and such that they very effectively cross-link B-cell receptors,
presentation by the MHC class II peptide presentation an important point since cross-linking of the B-cell recep-
pathway. In contrast, dendritic cells are indiscriminate tors is the key to triggering the signal (Fig. 4.20). The de-
phagocytosers. Consequently, B cells can specifically con- tails are as follows. The B-cell receptor has only a short
centrate antigen for presentation via the MHC class II cytoplasmic domain and therefore requires associated sig-
peptide presentation pathway, whereas much of what naling molecules to transmit a signal into the cell. These
dendritic cells present is from internalized cellular debris. B-cell-receptor-associated signaling molecules are re-
This distinction is important because activation of Th ferred to as Ig and Ig (Fig. 4.20). Together with the
cells requires cross-linking of their T-cell receptors. In- plasma membrane-associated IgM molecules (i.e., the
deed, there is a threshold number of T cell receptors that antibody class made by nave B cells), they form the
must be cross-linked for the Th cell to be activated. Since functional B-cell receptor. On the cytoplasmic side of the
B cells specifically concentrate and present specific anti- plasma membrane, Ig and Ig associate with the tyrosine
gen rather than other debris, they achieve a higher level of kinase Blk, Fyn, or Lyn. When the receptor complexes are
T-cell receptor cross-linking than other professional anti- cross-linked by the pathogen, the receptor-associated ty-
gen-presenting cells. This is especially consequential when rosine kinases cross-phosphorylate the cytoplasmic tails
low levels of antigen might be present. of Ig and Ig on neighboring cross-linked receptor com-
Th1 and Th2 cells have different fates following their plexes, thereby initiating the signal cascade. It is likely that
activation. Once CTLs and Th1 cells have been activated this B-cell receptor signaling process evolved to exploit
(and undergone clonal expansion), they migrate from the the regularly repeating surface features on viral and bacte-
lymph nodes, where activation occurs, to the infected rial pathogens. (As in the case of the heavy chains of the
sites. In contrast, Th2 cells stay in the lymphoid tissues, B-cell receptor, the and proteins of the T-cell receptor
where they transmit signals that help to activate B cells. have short cytoplasmic domains, consisting of only several
When Th1 cells reach the infected site, their commitment amino acids. Consequently, they too are not able to signal
to their Th1 cytokine profile may be strengthened or al- by themselves. However, the cytoplasmic domains of T-
tered by the particular cytokines they encounter there. Next, cell receptors are associated with the CD3 complex,
we consider the actual functions of activated Th cells. comprised of four different proteins, , , , and [zeta],
which transmit the signal when the T-cell receptor binds
Activation of B Cells to an MHC/peptide complex [Figure 4.19].)
One of the major roles for Th cells is to help B cells mature The signal provided by the activated Th cell to the B
into antibody-producing plasma cells. To appreciate why cell is referred to as the costimulatory signal (Fig. 4.21).
B cells might require help from T cells, recall how nave Yet the B cell itself must first play a key role before it can
T cells require two separate signals in order to be activated receive this signal. Recall that B-cell receptors bind their
Host Defenses and Viral Countermeasures 127

Clustering of antigen receptors allows receptor- Syk binds to doubly phosphorylated

associated kinases to phosphorylate the ITAMs ITAMs and is activated on binding




Blk, Fyn, or Lyn

Syk B cell Signals
Changes in gene
expression in nucleus

Figure4.20 B-cell receptors must be cross-linked by the regularly spaced epitopes on the pathogen in order to trans-
mit an activating signal. Since the B-cell receptor has only a short cytoplasmic domain, it requires associated signaling
molecules, referred to as Ig and Ig, to transmit a signal into the cell. On the cytoplasmic side of the plasma mem-
brane, Ig and Ig associate with the tyrosine kinase Blk, Fyn, or Lyn. When the receptor complexes are cross-linked
by the pathogen, the receptor-associated tyrosine kinases cross-phosphorylate the cytoplasmic tails of Ig and Ig on
neighboring cross-linked receptor complexes, thereby enabling the Syk tyrosine kinase to bind. Syk molecules, bound
to neighboring B-cell receptors, cross-phosphorylate each other, thereby triggering the signal cascade. The signal
reaches the nucleus, where it induces changes in gene expression, resulting in B-cell activation. Adapted from
P. Parham, The Immune System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.

cognate antigens, which they then internalize by endocy- transmits a signal back into the Th cell, causing it to re-
tosis and process for presentation by the MHC class II lease cytokines that, together with CD40 ligand and the
peptide presentation pathway. Only upon presenting its signal transmitted by the B-cell receptors, cause B cells to
cognate antigen by the MHC class II peptide presentation proliferate and mature into antibody-secreting plasma
pathway can the B cell be recognized by the particular cells. Important points to bear in mind are that Th cells
clone of activated Th cells. Importantly, while these B cells must be activated to deliver costimulatory signals to B cells
recognize the same pathogen as the activated Th cells, and Th cells can be activated only in response to an ongo-
recall that B-cell receptors recognize epitopes on whole ing infection.
antigens (pathogens), whereas T-cell receptors recognize These interactions of nave B cells with the pathogen
peptides presented in the context of MHC molecules. and with activated Th cells occur in the lymphoid organs.
Therefore, although the interacting Th cell and B cell are The pathogen and antigen-presenting dendritic cells are
specific for the same pathogen, they may well recognize carried to a lymph node primarily by the lymph that
different features on the pathogen. drains the infected tissue. In contrast, Th cells enter lym-
When the T-cell receptors on the Th cell bind to the phoid organs from the blood circulation, searching from
MHC class II/peptide complexes on the B cell, the Th cell one to another for their cognate antigen. Since activation
responds by synthesizing the CD40 ligand, which interacts of B cells requires prior activation of nave Th cells, which
with CD40 on the surface of the B cell. This interaction in turn was preceded by activation of dendritic cells, the
128 CHAP TER 4

Antigen binding to B cell receptor Helper TH2 cell delivers the second B cell proliferates and
delivers the first signal to the B cell signal via CD40 ligand and cytokines differentiates into plasma cells

Helper T cell


1 B cell 2

Figure4.21 B-cell activation requires two signals. The first is provided when the B-cell receptors are cross-linked by
the pathogen (left panel). The second or costimulatory signal is provided by activated Th cells (center panel). In order
for the B cell to receive the costimulatory signal, the B-cell receptors must first bind and internalize their cognate anti-
gen for processing and presentation by the MHC class II peptide presentation pathway. When the T-cell receptors on
the Th cell bind to the MHC class II/peptide complexes on the B cell, the Th cell responds by synthesizing the CD40 li-
gand, which interacts with CD40 on the surface of the B cell. This interaction transmits a signal back into the Th cell,
causing it to release cytokines that, together with CD40 ligand and the signal from the B-cell receptor, cause B cells
to proliferate and mature into antibody-secreting plasma cells (right panel). Adapted from P. Parham, The Immune
System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.

onset of antibody production is delayed until a week or so not be necessary. However, note that only the live virus
after infection. infection caused isotype switching to IgG production.
Here now is another of those irksome exceptions to Perhaps live virus infection causes signals to be generated
our neat generalizations. Experiments with T-cell-deficient by the innate immune system that might trigger the class
mice show that some pathogens can activate B cells to pro- switch. However, T-cell costimulatory signals may yet be
duce IgM and IgG antibodies without T-cell help. Antigens required for further antibody class switching and for so-
on these T-cell-independent pathogens are generally ar- matic hypermutation.
ranged in a very organized, highly repetitive fashion, which Does T-cell-independent activation of B cells play a
enables them to very extensively cross-link the B cell recep- role in natural virus infections in an immunocompetent
tor and thereby to deliver a strong activating signal to B host? The answer is not known with certainty. However,
cells, which can suffice to activate them. For example, some one might suspect that the answer is yes, in at least some
bacteria have highly organized repetitive polysaccharides instances, since T-cell-independent activation of B cells
on their surfaces, which extensively cross-link B-cell recep- could occur before T cells might be activated. Thus, T-cell-
tors, thereby activating B cells without T-cell help. independent activation of B cells could jump-start anti-
Experiments involving the mouse polyomavirus pro- body production, thereby helping to bring about earlier
vide an example of virus-induced T-cell-independent ac- control of the infection.
tivation of B cells. T-cell-deficient mice were infected with
live polyomavirus, or they were immunized with empty Activation of CTLs
virus-like particles (VLPs) that were assembled from pen- The activation of nave CD8 T cells to effector CTLs, like
tamers of the major polyomavirus capsid protein, VP1 the activation of Th cells and B cells, requires two signals.
(which was generated in recombinant insect cells). These Reviewing the activation of Th cells, we saw that Th cells
VLPs have the same antigenic structure as live virus. Other are activated when a dendritic cell delivers (i) a general
T-cell-deficient mice were immunized with an equal signal via B7, which is induced on the dendritic cell by
amount of soluble VP1. In each instance, IgM production cytokines and antigens at the infected site; and (ii) a spe-
was elicited in the T-cell-deficient mice. However, live vi- cific signal via MHC class II molecules, which present an-
rus and the VLPs elicited a 10-fold-stronger response than tigen from the invading pathogen. In the case of B-cell
soluble VP1molecules. These experimental results show activation, one of the activating signals is triggered when
that when the viral antigens are present in a highly orga- the B-cell receptors are cross-linked by their cognate anti-
nized, repetitive capsid structure, they more efficiently gen. The second or costimulatory signal is triggered by a
induce B cells to produce IgM, such that T-cell help may Th cell that recognizes its cognate antigen, as presented by
Host Defenses and Viral Countermeasures 129

the B cell in the context of an MHC class II molecule. The One signal is the antigen-specific activating signal, trans-
activation of nave CD8 T cells may result from several mitted via the interaction between the MHC class I/
variations on this two-signal scheme. However, in each peptide complex on the dendritic cell and the T-cell re-
case, activation requires the participation of an activated ceptor on the nave CD8 T cell. The other is a nonspecific
dendritic cell. costimulatory signal, transmitted by the interaction be-
Why the particular requirement for dendritic cells to tween B7 on the dendritic cell and CD28 on the nave
activate a nave CD8 T cell? The answer is in keeping with CD8 T cell (Fig. 4.22, left panel). However, the dendritic
an earlier maxim stating that the large energy outlay nec- cell will be ready to transmit a sufficiently strong costimu-
essary to generate CTLs and ensure their potency neces- latory signal only if it has been activated sufficiently at the
sitates ensuring that CTLs be activated only in response to infected site to express adequate levels of B7 to do so. The
an actual infection. Apropos that premise, recall that den- now-activated CTL then produces IL-2, driving its own
dritic cells can become antigen presenters only in response proliferation. Notice that this means by which a dendritic
to inflammatory cytokines and antigens at the infected cell activates a CTL is much the same as the way by which
site. Moreover, only dendritic cells, which are the most it activates a Th cell (see above).
potent of the professional antigen-presenting cells, can Even if the dendritic cell can offer only suboptimal co-
provide sufficient stimulation, via the several different stimulation via B7, it still can activate a nave CD8 T cell,
pathways described below, to activate a nave CD8 T cell but with the help of a Th cell. This can happen whether or
to become a CTL. not the CD4 T cell is already an activated Th cell. If the
Have you thought about how a dendritic cell might de- CD4 T cell is already an activated Th cell, then, upon bind-
liver an antigen-specific signal to a nave CD8 T cell? This ing to the dendritic cell, it induces the dendritic cell to
is not a trivial matter. A dendritic cell, as a professional increase its level of costimulators to levels that are suffi-
antigen-presenting cell, might be expected to present ex- cient to activate the CD8 T cell (Fig. 4.22, center panel).
ogenous peptides to the nave CD8 T cell via its MHC class Once again, the activated CTL produces IL-2, driving its
II peptide presentation pathway. But do you see a problem own proliferation. Note that the Th cell recognizes MHC
here? Recall that CD8 T cells specifically recognize pep- class II/peptide complexes on the dendritic cell, while the
tides presented on MHC class I molecules. So, what might CTL and its nave CD8 T cell precursor recognize MHC
be the way out from this conundrum? The answer is that class I/peptide complexes on the dendritic cell.
dendritic cells have the special ability to present exogenous Even if the CD4 T cell is still in its nave state, it still
peptides on both MHC class I and MHC class II mole- might be able to help activate the CTL. That can happen if
cules, another feature of dendritic cells making them the the nave CD8 T cell and the nave CD4 T cell should si-
most important of the professional antigen-presenting multaneously recognize the same antigen-presenting den-
cells. Thus, nave CD8 T cells can recognize their cognate dritic cell (Fig. 4.22, right panel). The nave CD4 T cell can
antigen when presented by the dendritic cells MHC class I then be activated by the dendritic cell to produce IL-2,
peptide presentation pathway. which in turn promotes the activation and proliferation of
The pathway by which dendritic cells present exogenous the nave CD8 T cell to generate a clone of mature CTLs.
antigenic peptides to nave CD8 T cells via MHC class I
molecules is known as cross-presentation. It is not yet Mechanism of Action of CTLs
well understood, but there is experimental evidence which Once a nave CD8 T cell has been activated and under-
suggests that some internalized proteins (in particular, gone clonal expansion, the resulting mature CTLs leave
those proteins that have formed immune complexes with the lymph node and enter the circulation in search of in-
antibodies) can somehow escape from the endosome/ fected target cells to attack and destroy. Circulating T cells
lysosome compartment into the cytosol, where they are ex- are directed to infected sites by particular chemokines,
posed to the ubiquitin/proteasome pathway, followed by i.e., small proteins that guide lymphocytes to sites of in-
transport of resulting peptides to the ER and presentation fection. These include macrophage inflammatory protein
by the MHC class I peptide presentation pathway. (What 1 (MIP-1; HIV uses the MIP-1 receptor, CCR5, for a
might be the consequences if other types of cells were able coreceptor [Chapter 21]) and IFN--inducible protein-10,
to present exogenous antigen via MHC class I molecules?) which are induced by IFN- in a variety of cell types at the
As noted above, the activation of CTLs, like that of infected site. When a CTL is thus directed to an infected
other immune effectors, requires two signals. Consequently, site, it scans the tissue for cells displaying MHC class I/
the most straightforward pathway by which a dendritic peptide complexes that its T-cell receptor might specifi-
cell can activate a nave CD8 T cell is for the dendritic cally recognize. Upon encountering such a target cell, the
cell to transmit both of these signals to the CD8 T cell. CTL then initiates the steps leading to its destruction.
130 CHAP TER 4

CD8 T cell CD4 T cell CD4 T cell

CD8 T cell CD8 T cell

CD28 CD28
CD8 CD4 CD8 IL-2
B7 B7


dendritic cell



Figure4.22 The activation of nave CD8 T cells to become effector CTLs requires two signals. (Left panels) A fully
activated dendritic cell can transmit both of these signals: (i) an antigen-specific activating signal transmitted by the
interaction between the MHC class I/peptide complex on the dendritic cell and the T-cell receptor on the nave CD8
T cell and (ii) a nonspecific costimulatory signal transmitted by the interaction between B7 on the dendritic cell and
CD28 on the nave CD8 T cell. However, this can occur only if the dendritic cell expresses levels of B7 that are suffi-
cient to transmit a strong enough costimulatory signal. (Center panels) Even if the dendritic cell can offer only subop-
timal costimulation via B7, it still can activate a nave CD8 T cell, but with the help of a CD4 T cell. If the CD4 T cell
already is an activated Th cell, then, upon binding to the dendritic cell, it induces the dendritic cell to increase its level
of costimulators to levels that are sufficient to activate the nave CD8 T cell. For example, note the upregulation of the
B7 costimulatory molecule. (Right panels) Even if the CD4 T cell is still in its nave state, it still might be able to help
activate the nave CD8 T cell, if the nave CD8 T cell and the nave CD4 T cell should simultaneously recognize the
same antigen-presenting dendritic cell. The nave CD4 T cell can then be activated by the dendritic cell to produce
IL-2, which in turn promotes the activation and proliferation of the nave CD8 T cell to generate a clone of mature
CTLs. Adapted from P. Parham, The Immune System, 2nd ed. (Garland Science, New York, NY, 2005), with permission.

The CTL kills the target cell by inducing it to undergo important advantages for the host. Necrotic cells eventu-
programmed cell death, also referred to as apoptosis (see ally suffer loss of membrane integrity and thus release po-
above). The apoptotic process is characterized by the frag- tentially damaging cellular contents into the tissues. Were
mentation of cellular DNA, breakup of the nucleus into a necrotic cell infected with a virus, infectious virus par-
smaller bodies, and finally, the shrinkage and breakup of ticles would very well be released as well. In contrast, the
the cell into membrane-enclosed vesicles called apoptotic apoptotic process rapidly degrades viral nucleic acids, re-
bodies (Fig. 4.23). sulting in the rapid cessation of virus replication in the
As noted earlier, apoptosis is a normal process in the cell. Moreover, since the apoptotic cell does not lyse, no
host, in which it serves several important functions not intracellular virus that might be present is released. In ad-
necessarily related to immune defenses. Regardless, the dition, no potentially damaging cellular constituents are
CTL-mediated destruction of virus-infected target cells released into the tissues. Also, the apoptotic bodies are
by apoptosis rather than by necrosis (the type of cell death readily phagocytosed by tissue macrophages and cleared
resulting from acute cellular injury or trauma) has several from the infected site. Finally, and importantly, apoptotic
Host Defenses and Viral Countermeasures 131

membrane, and the granules contents are released onto

the target cell. The free perforin then may assemble mem-
brane attack complexes on the target cell membrane,
thereby creating pores that enable the granzyme to enter
the target cell, where it induces apoptosis. (Other descrip-
DNA fragmentation tions of this process have the target cell endocytose the
Chromosome condensation perforin and granzyme that were released by the CTL.
Within the target cell, the perforin molecules then create
pores in the membranes of the endocytic vesicles. This
allows the granzyme B to enter the cytoplasm, where it
induces the sequence of reactions leading to apoptosis.)
The second mechanism for inducing apoptosis is the
Fragmentation of nucleus well-characterized pathway involving the cell surface pro-
teins Fas and FasL. FasL, on the surface of the CTL, binds
to Fas on the surface of the target cell, thereby transmit-
ting a signal into the target cell that triggers apoptosis.
Why have two mechanisms by which CTLs can induce
apoptosis? Here are several possible reasons. First, not all
cells express Fas, so the Fas-mediated process cannot
Fragmentation of cell suffice to kill all virus-infected target cells. Second, as
Engulfment by macrophage demonstrated by experiments in which perforin and Fas
knockout mice were infected with influenza virus, opti-
mal clearance of the virus occurs when both perforin and
Fas can act. Thus, the two mechanisms have an additive
effect, at least in some cases. Third, the perforin and Fas
pathways generally serve different primary purposes. The
Figure4.23 CTLs kill target cells by inducing them to undergo perforin pathway is a lot quicker than the Fas pathway.
programmed cell death, also referred to as apoptosis. The apoptotic Perhaps that is why it is the major pathway for killing virus-
process is characterized by several distinct stages, beginning with the
infected cells. In contrast, the Fas pathway is the primary
fragmentation and compressing of cellular DNA, which is seen in dis-
tinct masses at the nuclear envelope. The cytoplasm then condenses, means for disposing of self-reactive T cells during negative
and the nucleus breaks into smaller bodies. In the end, the cell breaks selection in the thymus (see Self Tolerance, below). This
up into membrane-enclosed vesicles called apoptotic bodies, which distinction raises the following thought. Once an infection
may be engulfed and destroyed by macrophages. Adapted from
has been cleared, it is important to eliminate the expanding
G. M. Cooper and R. E. Hausman, The Cell: a Molecular Approach,
4th ed. (ASM Press, Washington, DC, 2007), with permission. population of activated, cytokine-secreting T cells. Thus, just
as apoptosis, as mediated by the Fas pathway, is the primary
means for disposing of self-reactive T cells during negative
selection in the thymus, apoptosis induced by the Fas-FasL
bodies also are taken up by dendritic cells, which then interaction between lymphocytes is the main means for
process and present the antigen within them to T cells. If disposing of unwanted lymphocytes after an infection has
apoptosis is blocked in infected cells, then antigen acquisi- been resolved. Individuals with a mutant FAS gene have a
tion by dendritic cells is impaired, and consequently, condition known as autoimmune lymphoproliferative
antigen presentation to T cells is impaired as well. syndrome. They cannot remove autoreactive T cells, nor
CTLs can induce apoptosis by two distinct mecha- can they control the size of their lymphocyte population.
nisms. Actually, these mechanisms are the same as those
used by NK cells (see above). The first of these mecha- T Cells and Antiviral Cytokines
nisms, referred to as the granule exocytosis pathway, is Despite the fact that Th cells and CTLs have crucially im-
dependent on the protein perforin and an enzyme called portant roles in helping to activate B cells and CTLs and
granzyme B. Perforin is a close relative of complement in killing virus-infected cells, these cells have yet other
protein C9, with which it shares the ability to poke holes vital functions to perform. In particular, when both CD4
in cell membranes. These proteins are stored in granules and CD8 T cells are activated, they produce large quanti-
within the activated CTL. When the CTL recognizes and ties of antiviral cytokines, including IFN- and TNF-.
binds to a target, the granules migrate to the CTLs plasma Production of these cytokines by activated T cells very
132 CHAP TER 4

likely contributes to the resolution of at least some virus the virus to generate proteins that are resistant to the
infections. It is probably important in this regard that ubiquitin/proteasome pathway. (Had this possibility oc-
IFN- up-regulates expression of MHC molecules and curred to you?) This is a particularly effective strategy in
other antiviral proteins, including PKR, 2-5-oligo(A) the case of EBV, a member of the herpesvirus family which,
synthetase, ADAR, and NO synthetase (see above). Also, like other herpesviruses, generally establishes latent infec-
since T-cell expression of IFN- and TNF- is dependent tions. While in the latent state, herpesviruses express very
on contact with antigen, production of these cytokines few proteins. (Note that this is in itself a CTL evasion
and their potentially damaging impact are thereby re- strategy.) In fact, the EBV nuclear antigen-1may be the
stricted to the infected site. only viral protein expressed in some latently infected B cells,
Another important feature of cytokines such as IFN- and it is the only viral protein expressed in EBV-associated
and TNF- is that their antiviral effects can occur in the malignancies (see Chapter 18). Yet EBV nuclear antigen-1
absence of target cell destruction. This may be particu- is not presented by the MHC class I peptide presentation
larly important in instances where vital organs might be pathway because it contains an amino acid sequence that
infected and excess tissue damage must be avoided. For inhibits the host proteasome.
example, IFN- and TNF-, as expressed by CTLs, were Latency is a central feature of the life cycles of all her-
shown to inactivate HBV within hepatocytes by a noncy- pesviruses. Yet for latency to contribute to the success of
tolytic process. these viruses in Nature, it must be followed by episodes of
So, a message to take away is that T cells act to resolve reactivation so that these viruses can be transmitted to
virus infections by means of the cytokines they secrete as susceptible individuals in the population. During these
well as by their cytolytic mechanisms. The effectiveness of reactivation episodes, herpesviruses need to express the
these respective defensive stratagems depends, as you full range of their proteins, and they also must contend
might have supposed, on the particular virus and the with primed immune systems, since memory cells are in
particular tissues that are infected. Also, as you might ex- place in latently infected individuals to jump-start adap-
pect, the optimum immune response in most instances tive immune responses when reactivation does occur.
requires the combined action of both cytokines and cy- Therefore, as we see immediately below, the herpesviruses
tolytic mechanisms. display a remarkable array of countermeasures against cell-
mediated immunity.
Viral Evasion of Cell-Mediated Immunity Several herpesviruses block TAP-mediated peptide
I hope that I have succeeded in conveying a sense that we transport into the ER. One such example is provided by
possess cell-mediated immune defenses of astonishing so- an HSV protein, ICP47, which binds to the cytoplasmic
phistication and potency. Yet the same evolutionary forces side of the Tap transporter (Fig. 4.24). Another herpesvi-
responsible for these immune defenses also have been at rus, HCMV, encodes a protein, US6, which blocks the Tap
work in the service of viruses. Moreover, as we noted transporter from the luminal side of the ER membrane.
above, viruses are able to evolve much faster than their The herpesviruses, as well as members of other virus
hosts. The results of virus evolution are perhaps nowhere families, generate proteins that either directly or indirectly
more impressive than in the range of countermeasures, inhibit the ability of MHC class I molecules to present an-
discussed below, which viruses have adapted to cope with tigenic peptides. For example, the HCMV US3 protein and
cell-mediated immune defenses. Indeed, virtually every the adenovirus E3 19K protein each physically associate
stage of the MHC class I and MHC class II peptide pre- with peptide-loaded MHC class I molecules in the ER,
sentation pathways is impeded by one or another virus. (It thereby preventing their transport to the cell surface. An-
is historically interesting that many of the steps of antigen other stratagem is displayed by the HCMV US11 and US2
presentation via MHC molecules actually were discovered proteins, which cause MHC class I proteins to be shed from
in the course of investigating virus mechanisms for evad- the ER into the cytoplasm, where they are then degraded.
ing adaptive immunity.) In addition, viruses also have (The interactions of US11 and US2 with MHC class I mol-
evolved measures that obstruct apoptosis as a host antiviral ecules in the ER apparently target the MHC class I mole-
defense. Below, we consider but a few examples to acquire a cules to the normal cellular pathway for the degradation
further appreciation of the seemingly unlimited resource- of misfolded proteins. That is, US11- and US2-bound
fulness of viruses to find ways to evade host defenses. MHC class I molecules are removed from the ER via Sec61
pores and degraded by cytosolic proteasomes.)
Inhibition of Antigen Presentation to CTLs If you have been counting, you may have noticed that
A rather direct strategy for avoiding antigen presentation HCMV encodes at least four inhibitors of the MHC class I
via the MHC class I peptide presentation pathway is for peptide presentation pathway: US2, US3, US6, and US11.
Host Defenses and Viral Countermeasures 133

Cytotoxic T lymphocyte

T-cell receptor

Infected cell HIV Nef MCMV gp48

Viral and
cellular proteins

MCMV m152
Proteasome Adenovirus E3 19K

MHC heavy chain Endoplasmic reticulum
Peptides HCMV US6


P 2-microglobulin

P MHC I Nucleus proteasome

Adenovirus EIA

Figure4.24 Viruses have evolved stratagems that impede the MHC class I peptide presentation pathway at virtually
every stage, including the processing of viral proteins in the proteasome and the transport of the resulting peptides to
the ER, the transcription of MHC class I proteins, and their transport to, and retention at, the plasma membrane. Spe-
cific viral gene products are highlighted in the yellow boxes. Steps in the antigen presentation pathway that are blocked
by particular viral gene products are indicated by blunt-ended red bars, whereas steps that they stimulate are indicated
by green arrows. Adapted from S. J. Flint, L. W. Enquist, R. M. Krug, V. R. Racaniello, and A. M. Skalka, Principles of
Virology: Molecular Biology, Pathogenesis, and Control, 2nd ed. (ASM Press, Washington, DC, 2003), with permission.

One might ask why such apparent redundancy of func- MCMV protein, m152, causes MHC class I/peptide com-
tion is necessary. A straightforward answer is that the plexes to be retained in the ER-Golgi intermediate com-
multiple gene products likely act additively to impair MHC partment. In addition, m152 is thus far unique in that it
class I function. Also, individual viral proteins might work also interferes with the expression of the cellular ligand
better against different MHC class I isoforms. Regardless, for the NK cell receptor, whose engagement provides acti-
the evolution of these multiple virus countermeasures vating signals to NK cells. (We noted above that HCMV,
against MHC class I-mediated defenses strongly implies the human cytomegalovirus, also resists attack by NK cells,
that these defenses would otherwise be a singularly impor- in that instance by expressing an MHC class I homolog
tant obstacle to these viruses. that acts as a decoy for NK cell inhibitory receptors.)
The mouse cytomegalovirus (MCMV) likewise encodes HIV expresses two proteins that ultimately cause the
multiple proteins that block the MHC class I peptide pre- down-regulation of MHC class I molecules at the cell sur-
sentation pathway. One of these, gp48, binds to MHC face. One of these proteins, Vpu, accomplishes this by
class I complexes and redirects them to the endocytic route destabilizing newly synthesized MHC class I molecules in
for rapid proteolytic destruction in lysosomes. Another the ER, thereby triggering their destruction by proteasomes.
134 CHAP TER 4

The other HIV protein, Nef, interrupts the normal recy- unaffected and thus be able to present exogenous HCMV
cling of MHC class I chains from the plasma membrane proteins via the MHC class II pathway. How then is the
back to the ER and instead diverts them to lysosomes, virus benefited by its ability to shut down antigen presen-
where they are degraded. tation via MHC class II molecules during productive in-
The HIV Nef protein has another effect, which pro- fection? The answer to this question is as follows. HCMV
vides one of many examples of how diabolical a virus is indeed does productively infect macrophages and en-
HIV. Nef selectively down-regulates HLA-A and HLA-B, dothelial cells, and these cells are an important reservoir
while not affecting HLA-C. Although HLA-C proteins can for the virus. Macrophages, as we know, are professional
present viral peptides to CTLs, they are the predominant antigen-presenting cells. As for endothelial cells, they ac-
molecules that transmit inhibitory signals to NK cells, in tually acquire a limited capacity to present antigen via the
that way preventing killing of a target cell by an NK cell. MHC class II pathway and to express B7. The limited ca-
The net result is that the MHC class I peptide-presentation pacity of endothelial and epithelial cells to express MHC
pathway is sufficiently compromised in HIV-infected cells class II molecules and to present antigen to CD4 Th cells
to impair attack by CTLs, while the unimpaired expres- is induced by IFN-, which is produced by the innate im-
sion of HLA-C enables HIV-infected cells to also resist at- mune response to the infection. In addition, during the
tack by NK cells. normal course of the HCMV replication cycle, some en-
In the last of these selected examples, the adenovirus dogenous viral proteins do cycle through endosomal com-
E1A protein and the HIV Tat protein each block expres- partments where MHC class II peptide loading occurs.
sion of MHC class I genes by blocking transcription from Consequently, by impairing MHC class II-mediated pre-
the MHC class I promoter. (See Tat Revisited in Chap- sentation of endogenous viral peptides, HCMV prevents
ter 21 for more of the several means by which the remark- infected macrophages and endothelial cells from being
able HIV Tat protein impairs antigen presentation by detected by CD4 Th cells. This provides an important
HIV-infected cells.) advantage, since CD4 Th cells that recognize HCMV-
infected macrophages and endothelial cells can respond
Inhibition of Antigen Presentation by producing antiviral cytokines. Moreover, a subset of
to Helper T Cells CD4 T cells actually have CTL function (Box4.12).
Considering that the recognition of MHC class II/peptide Since (i) the ability of endothelial cells to express MHC
complexes by Th cells is essential for the activation of both class II molecules and present antigen to CD4 Th cells is
antibody and cell-mediated antiviral immune responses, induced by IFN-, and (ii) IFN- transmits signals via the
you might expect that some viruses have evolved ways to JAK/STAT signaling pathway, it would be advantageous to
impair the MHC class II peptide presentation pathway. a virus to be able to short-circuit the JAK/STAT pathway.
Also, since as many as one-half of the 80 or more herpes- Members of the herpesvirus and adenovirus families and
virus genes play a role in evading host immune defenses other viruses indeed inhibit the JAK/STAT pathway, thereby
(Chapter 18), you might expect that some of the herpesvi- blocking the effects of IFN-, as well as of IFN-, on in-
rus immune evasion genes might serve to impede the fected cells. These viral countermeasures can be par-
MHC class II peptide presentation pathway. And indeed, ticularly important when endothelial and epithelial
that is the case. For example, the HCMV US2 protein, cells are infected, since MHC class II expression can be
which causes MHC class I proteins to be shed from the ER up-regulated in these cells from low basal levels by cyto
into the cytoplasm to be degraded, likewise targets MHC kines produced at the infected site.
class II chains for degradation. The HCMV US3 protein,
which retains MHC class I proteins in the ER, also destabi- Inh